LING073 - User contributions [en]

Kaingang and Portuguese/Final Project

2019-05-15T04:07:43Z

Drosset1: /* Specs for rule-based */

==The project: RBMT or NMT?==
===Background===
Kaingang is a language spoken in the southern Brazil region. It has roughly 20,000 speakers, many of whom are bilingual with Portuguese. Kaingang is one the most well-documented languages of indigenous peoples of Brazil; however, in our project, we were only able to use a [https://www.sil.org/resources/archives/42876 Kaingang-Portuguese dictionary] and a [https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf Kaingang new testament bible] (the other resources were written in languages we didn't read).
===Idea===
We wanted to find out how well a neural machine translation could work with a language that didn't have millions of parallel lines (in fact, we have ~10,000). We also wanted to compare it to our rule-based transducer, which we've been making additions to since the end of the semester.

===Motivation===
This is the only (at least easily accessible when searching) machine translation available between Kaingang and another language. The motivation behind our project was to possibly allow the Kaingang people to not have to choose between teaching Kaingang or Portuguese to later generations. We hoped (if at some point one of the translators was state-of-the-art) any need for Portuguese could be accessed through the translator, which would be an incentive to keep the language Kaingang alive.

==What we did==
* Added many more features (such as much larger coverage, many more words added to the bilingual dictionary, new disambiguation rules, and 2 new structural transfer rules) and improved performance of rule-based transducer
* Trained a neural network with 75% of our parallel corpus and tested with 25% using OpenNMT
* Documented ambiguity in Kaingang
* Created parallel corpora for the whole new testament between Kaingang and Portuguese!

==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]

[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation Metric==
Our new evaluation accounts for the fact that words in the Apertium translation sentences has # and * symbols. The accuracy script looks something like this:
correct_word_percentage = []
for i in range(len(true_translations)):
correct_word_count = 0
for word in prediction[i].split():
if word.strip("\n").strip("*").strip("#") in true_translations[i]:
correct_word_count += 1
correct_word_percentage.append(correct_word_percentage/len(prediction[i].split()))
average_accuracy = sum(correct_word_percentage)/len(correct_word_percentage)
print(average_accuracy)
===Accuracy===
Initial RBMT accuracy (actually higher than poster because of new evaluation metric): 6.78%

Final RBMT accuracy: 9.91%

Initial NMT accuracy: 21.4%

Final NMT accuracy: (currently training on new larger corpus, so no final accuracy yet)

==Rule-based coverage==

* Coverage over new testament:
$ aq-covtest ../ling073-kgp-corpus/kgp.corpus.large.txt kgp.automorf.bin
Number of tokenised words in the corpus: 406362
Coverage: 92.93%
Top unknown words in the corpus:
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
270 nĩnh
269 tũꞌ
Translation time: 1.422363042831421 seconds
* Coverage over dictionary sentences:
$ aq-covtest ../ling073-kgp-por-corpus/kgp.sentences.txt kgp.automorf.bin
Number of tokenised words in the corpus: 32659
Coverage: 92.72%
Top unknown words in the corpus:
30 nĩm
27 vég
26 isóg
24 jagy
23 prũ
23 tãg
22 kyr
20 mỹnh
20 régre
18 pir
18 féj
17 jã
17 Kusa
17 nĩgé
16 kãtĩ
16 jamã
15 venh
15 gãnh
15 jakré
14 nug
Translation time: 0.1878523826599121 seconds
===Bilingual Coverage===
* Number of times "*" showed up in translation of the bible initially: 97380
* Number of times "*" shows up in the translation of the bible now: 59835

==Further work==
===Documented ambiguity in Kaingang===
We haven't employed rules for the vast majority of these but we compiled a list of all ambiguous words in the translator!
ag - (pr pes) eles / (pr dem) os
e - (mod) muito / (v iv) fazer, causar
ẽmĩ - (n) pao de milho / (v iv) tatear, apalpar
fa - (n) perna, planta, amargo / (v tv sg) lavar roupa
fag - (pr pes) elas / (pr dem) as
fi - (pr pes) ela / (pr dem) a
ge - (mod) entao, tambem / (n) semelhante / (v iv) fazer igual
gé - (mod) tambem / (v tv pl) levar, carregar
gunhgunh - (n) picao preto / (v tv pl) fincar
han - (v iv sg) sarar, melhorar / (v tv sg) fazer
he - (ij) sim / (v tv) dizer
hur - (o) ja / (v iv sg) caimbra
hỹn - (o) de certo, provavelmente / (int) onde?
jãn - (v tv sg) rasgar / (v iv sg) rezar, cantar hinos religiosos
jãvãnh - (mod) recusar, nao saber fazer / (v tv sg) esperar, procurar
jé - (su) futuro / (n) reza, hino
jẽn - (v i) comer / (v tv) dar de comer
jo - (conj) entao / (cir) antes, na frente
jũn - (n) palmito / (v tv) embrabecer
kafãn - (v tv sg) descascar milho / (v iv sg) emparelhar, responder, discutir
ke - (n) sobra, resto / (v iv) futuro, fazer, dizer
ker - (o) cuidado / (n) presente de comida, cama de barbante de urtiga
kre - (n) balaio, coxa / (v tv pl) cortar, ceifar
kren - (mod) quase / (v iv sg) escapar / (v tv sg) perder
krenkren - (v iv pl) escapar / (v tv pl) perder
kur - (o) ligeiro, rapido! / (n) roupa, vestido
mé - (mod) fazer diariamente, ligeiro, gostar de fazer / (n) carneiro
mẽ - (mod) muito, ligeiro / (v tv sg) cheirar, escutar, sentir, tocar
mro - (v iv sg) tomar banho / (v tv sg) por de molho
mũ - (a) fazendo, narrativo, acao unica, consequencia / (v iv pl) andar
mỹ - (cir) para / (su) pergunta
mỹr - (o) verdadeiro, e certo / (conj) ?
ne - (su) originador / (v tv sg) enterrar / (n) enterrado / de / (int) o que?
nẽ - (a) sera que? / (v tv) cobrir encostado
nig - (n) lagoa, poco / (v tv sg) dar pontape
nĩ - (a) na situacao de, sentado, ter, ter a obrigacao, condicao / (n) carne / (v iv sg) sentar
nỹ - (a) deitado / (su) topico na pergunta / (n) mae, irma da mae / (v iv sg) deitar-se
nỹtĩ - (a) sendo, ter / (v iv pl) existir
pẽ - (mod) muito, o verdadeiro / (n) braco
ra - (cir) para, na direcao de, apesar de / (n) queixo / (a) faca ja!
rã - (cir) perto, por baixo de / (n) sol, maduro / (v iv sg) entrar, comecar, estar perto de
rán - (n) encosta, declive, barranco, subida, perau / (v tv sg) escrever
rãké - (v iv) chegando tarde / tarde ?
rãnhrãj - (n) trabalho / (v tv pl) trabalhar
re - (n) grama, campo, fora de casa / (v iv) encher-se de bebida / (v iv pl) descer
sa - (a) pendurado / (n) sal, cachoeira / (v iv) pendurado / (v tv) pendurar, pregar
se - (n) quati / (v tv sg) amarrar, prender / (ij) puxa!
tãnh - (n) morador, dono / (v iv) fazer forte, habituar-se
ti - (pr pes) ele / (pr dem) o
tógfĩn - (n) gaviao / (v tv sg) atar, amarrar
tỹ - (cir) com, por / (erg) / (ex) / (top)
vãn - (n) taquara / (v tv sg) carregar coisa comprida
vãnh - (mod) nao querer ou poder fazer algo / (n) capoeira, mato, plantas curtas
ve - (n) aparencia, natureza, irma de um homem, primeiro / (v iv sg) parecer ser / (v tv sg) ver, enxergar
vo - (v iv) barulho de andar no mato, coracao batendo / (n) macuco
vỹ - (su) topico / (o) sera que e? / (voc) homem! moco!

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T04:07:07Z

Drosset1: /* Word Error Rate */

==The project: RBMT or NMT?==
===Background===
Kaingang is a language spoken in the southern Brazil region. It has roughly 20,000 speakers, many of whom are bilingual with Portuguese. Kaingang is one the most well-documented languages of indigenous peoples of Brazil; however, in our project, we were only able to use a [https://www.sil.org/resources/archives/42876 Kaingang-Portuguese dictionary] and a [https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf Kaingang new testament bible] (the other resources were written in languages we didn't read).
===Idea===
We wanted to find out how well a neural machine translation could work with a language that didn't have millions of parallel lines (in fact, we have ~10,000). We also wanted to compare it to our rule-based transducer, which we've been making additions to since the end of the semester.

===Motivation===
This is the only (at least easily accessible when searching) machine translation available between Kaingang and another language. The motivation behind our project was to possibly allow the Kaingang people to not have to choose between teaching Kaingang or Portuguese to later generations. We hoped (if at some point one of the translators was state-of-the-art) any need for Portuguese could be accessed through the translator, which would be an incentive to keep the language Kaingang alive.

==What we did==
* Added many more features (such as much larger coverage, many more words added to the bilingual dictionary, new disambiguation rules, and 2 new structural transfer rules) and improved performance of rule-based transducer
* Trained a neural network with 75% of our parallel corpus and tested with 25% using OpenNMT
* Documented ambiguity in Kaingang
* Created parallel corpora for the whole new testament between Kaingang and Portuguese!

==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]

[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation Metric==
Our new evaluation accounts for the fact that words in the Apertium translation sentences has # and * symbols. The accuracy script looks something like this:
correct_word_percentage = []
for i in range(len(true_translations)):
correct_word_count = 0
for word in prediction[i].split():
if word.strip("\n").strip("*").strip("#") in true_translations[i]:
correct_word_count += 1
correct_word_percentage.append(correct_word_percentage/len(prediction[i].split()))
average_accuracy = sum(correct_word_percentage)/len(correct_word_percentage)
print(average_accuracy)
===Accuracy===
Initial RBMT accuracy (actually higher than poster because of new evaluation metric): 6.78%

Final RBMT accuracy: 9.91%

Initial NMT accuracy: 21.4%

Final NMT accuracy: (currently training on new larger corpus, so no final accuracy yet)

==Specs for rule-based==
===Coverage===

* Coverage over new testament:
$ aq-covtest ../ling073-kgp-corpus/kgp.corpus.large.txt kgp.automorf.bin
Number of tokenised words in the corpus: 406362
Coverage: 92.93%
Top unknown words in the corpus:
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
270 nĩnh
269 tũꞌ
Translation time: 1.422363042831421 seconds
* Coverage over dictionary sentences:
$ aq-covtest ../ling073-kgp-por-corpus/kgp.sentences.txt kgp.automorf.bin
Number of tokenised words in the corpus: 32659
Coverage: 92.72%
Top unknown words in the corpus:
30 nĩm
27 vég
26 isóg
24 jagy
23 prũ
23 tãg
22 kyr
20 mỹnh
20 régre
18 pir
18 féj
17 jã
17 Kusa
17 nĩgé
16 kãtĩ
16 jamã
15 venh
15 gãnh
15 jakré
14 nug
Translation time: 0.1878523826599121 seconds
===Bilingual Coverage===
* Number of times "*" showed up in translation of the bible initially: 97380
* Number of times "*" shows up in the translation of the bible now: 59835
==Further work==
===Documented ambiguity in Kaingang===
We haven't employed rules for the vast majority of these but we compiled a list of all ambiguous words in the translator!
ag - (pr pes) eles / (pr dem) os
e - (mod) muito / (v iv) fazer, causar
ẽmĩ - (n) pao de milho / (v iv) tatear, apalpar
fa - (n) perna, planta, amargo / (v tv sg) lavar roupa
fag - (pr pes) elas / (pr dem) as
fi - (pr pes) ela / (pr dem) a
ge - (mod) entao, tambem / (n) semelhante / (v iv) fazer igual
gé - (mod) tambem / (v tv pl) levar, carregar
gunhgunh - (n) picao preto / (v tv pl) fincar
han - (v iv sg) sarar, melhorar / (v tv sg) fazer
he - (ij) sim / (v tv) dizer
hur - (o) ja / (v iv sg) caimbra
hỹn - (o) de certo, provavelmente / (int) onde?
jãn - (v tv sg) rasgar / (v iv sg) rezar, cantar hinos religiosos
jãvãnh - (mod) recusar, nao saber fazer / (v tv sg) esperar, procurar
jé - (su) futuro / (n) reza, hino
jẽn - (v i) comer / (v tv) dar de comer
jo - (conj) entao / (cir) antes, na frente
jũn - (n) palmito / (v tv) embrabecer
kafãn - (v tv sg) descascar milho / (v iv sg) emparelhar, responder, discutir
ke - (n) sobra, resto / (v iv) futuro, fazer, dizer
ker - (o) cuidado / (n) presente de comida, cama de barbante de urtiga
kre - (n) balaio, coxa / (v tv pl) cortar, ceifar
kren - (mod) quase / (v iv sg) escapar / (v tv sg) perder
krenkren - (v iv pl) escapar / (v tv pl) perder
kur - (o) ligeiro, rapido! / (n) roupa, vestido
mé - (mod) fazer diariamente, ligeiro, gostar de fazer / (n) carneiro
mẽ - (mod) muito, ligeiro / (v tv sg) cheirar, escutar, sentir, tocar
mro - (v iv sg) tomar banho / (v tv sg) por de molho
mũ - (a) fazendo, narrativo, acao unica, consequencia / (v iv pl) andar
mỹ - (cir) para / (su) pergunta
mỹr - (o) verdadeiro, e certo / (conj) ?
ne - (su) originador / (v tv sg) enterrar / (n) enterrado / de / (int) o que?
nẽ - (a) sera que? / (v tv) cobrir encostado
nig - (n) lagoa, poco / (v tv sg) dar pontape
nĩ - (a) na situacao de, sentado, ter, ter a obrigacao, condicao / (n) carne / (v iv sg) sentar
nỹ - (a) deitado / (su) topico na pergunta / (n) mae, irma da mae / (v iv sg) deitar-se
nỹtĩ - (a) sendo, ter / (v iv pl) existir
pẽ - (mod) muito, o verdadeiro / (n) braco
ra - (cir) para, na direcao de, apesar de / (n) queixo / (a) faca ja!
rã - (cir) perto, por baixo de / (n) sol, maduro / (v iv sg) entrar, comecar, estar perto de
rán - (n) encosta, declive, barranco, subida, perau / (v tv sg) escrever
rãké - (v iv) chegando tarde / tarde ?
rãnhrãj - (n) trabalho / (v tv pl) trabalhar
re - (n) grama, campo, fora de casa / (v iv) encher-se de bebida / (v iv pl) descer
sa - (a) pendurado / (n) sal, cachoeira / (v iv) pendurado / (v tv) pendurar, pregar
se - (n) quati / (v tv sg) amarrar, prender / (ij) puxa!
tãnh - (n) morador, dono / (v iv) fazer forte, habituar-se
ti - (pr pes) ele / (pr dem) o
tógfĩn - (n) gaviao / (v tv sg) atar, amarrar
tỹ - (cir) com, por / (erg) / (ex) / (top)
vãn - (n) taquara / (v tv sg) carregar coisa comprida
vãnh - (mod) nao querer ou poder fazer algo / (n) capoeira, mato, plantas curtas
ve - (n) aparencia, natureza, irma de um homem, primeiro / (v iv sg) parecer ser / (v tv sg) ver, enxergar
vo - (v iv) barulho de andar no mato, coracao batendo / (n) macuco
vỹ - (su) topico / (o) sera que e? / (voc) homem! moco!

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T04:06:58Z

Drosset1: /* Specs for rule-based */

==The project: RBMT or NMT?==
===Background===
Kaingang is a language spoken in the southern Brazil region. It has roughly 20,000 speakers, many of whom are bilingual with Portuguese. Kaingang is one the most well-documented languages of indigenous peoples of Brazil; however, in our project, we were only able to use a [https://www.sil.org/resources/archives/42876 Kaingang-Portuguese dictionary] and a [https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf Kaingang new testament bible] (the other resources were written in languages we didn't read).
===Idea===
We wanted to find out how well a neural machine translation could work with a language that didn't have millions of parallel lines (in fact, we have ~10,000). We also wanted to compare it to our rule-based transducer, which we've been making additions to since the end of the semester.

===Motivation===
This is the only (at least easily accessible when searching) machine translation available between Kaingang and another language. The motivation behind our project was to possibly allow the Kaingang people to not have to choose between teaching Kaingang or Portuguese to later generations. We hoped (if at some point one of the translators was state-of-the-art) any need for Portuguese could be accessed through the translator, which would be an incentive to keep the language Kaingang alive.

==What we did==
* Added many more features (such as much larger coverage, many more words added to the bilingual dictionary, new disambiguation rules, and 2 new structural transfer rules) and improved performance of rule-based transducer
* Trained a neural network with 75% of our parallel corpus and tested with 25% using OpenNMT
* Documented ambiguity in Kaingang
* Created parallel corpora for the whole new testament between Kaingang and Portuguese!

==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]

[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation Metric==
Our new evaluation accounts for the fact that words in the Apertium translation sentences has # and * symbols. The accuracy script looks something like this:
correct_word_percentage = []
for i in range(len(true_translations)):
correct_word_count = 0
for word in prediction[i].split():
if word.strip("\n").strip("*").strip("#") in true_translations[i]:
correct_word_count += 1
correct_word_percentage.append(correct_word_percentage/len(prediction[i].split()))
average_accuracy = sum(correct_word_percentage)/len(correct_word_percentage)
print(average_accuracy)
===Accuracy===
Initial RBMT accuracy (actually higher than poster because of new evaluation metric): 6.78%

Final RBMT accuracy: 9.91%

Initial NMT accuracy: 21.4%

Final NMT accuracy: (currently training on new larger corpus, so no final accuracy yet)

==Specs for rule-based==
===Coverage===

* Coverage over new testament:
$ aq-covtest ../ling073-kgp-corpus/kgp.corpus.large.txt kgp.automorf.bin
Number of tokenised words in the corpus: 406362
Coverage: 92.93%
Top unknown words in the corpus:
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
270 nĩnh
269 tũꞌ
Translation time: 1.422363042831421 seconds
* Coverage over dictionary sentences:
$ aq-covtest ../ling073-kgp-por-corpus/kgp.sentences.txt kgp.automorf.bin
Number of tokenised words in the corpus: 32659
Coverage: 92.72%
Top unknown words in the corpus:
30 nĩm
27 vég
26 isóg
24 jagy
23 prũ
23 tãg
22 kyr
20 mỹnh
20 régre
18 pir
18 féj
17 jã
17 Kusa
17 nĩgé
16 kãtĩ
16 jamã
15 venh
15 gãnh
15 jakré
14 nug
Translation time: 0.1878523826599121 seconds
===Bilingual Coverage===
* Number of times "*" showed up in translation of the bible initially: 97380
* Number of times "*" shows up in the translation of the bible now: 59835
===Word Error Rate===

==Further work==
===Documented ambiguity in Kaingang===
We haven't employed rules for the vast majority of these but we compiled a list of all ambiguous words in the translator!
ag - (pr pes) eles / (pr dem) os
e - (mod) muito / (v iv) fazer, causar
ẽmĩ - (n) pao de milho / (v iv) tatear, apalpar
fa - (n) perna, planta, amargo / (v tv sg) lavar roupa
fag - (pr pes) elas / (pr dem) as
fi - (pr pes) ela / (pr dem) a
ge - (mod) entao, tambem / (n) semelhante / (v iv) fazer igual
gé - (mod) tambem / (v tv pl) levar, carregar
gunhgunh - (n) picao preto / (v tv pl) fincar
han - (v iv sg) sarar, melhorar / (v tv sg) fazer
he - (ij) sim / (v tv) dizer
hur - (o) ja / (v iv sg) caimbra
hỹn - (o) de certo, provavelmente / (int) onde?
jãn - (v tv sg) rasgar / (v iv sg) rezar, cantar hinos religiosos
jãvãnh - (mod) recusar, nao saber fazer / (v tv sg) esperar, procurar
jé - (su) futuro / (n) reza, hino
jẽn - (v i) comer / (v tv) dar de comer
jo - (conj) entao / (cir) antes, na frente
jũn - (n) palmito / (v tv) embrabecer
kafãn - (v tv sg) descascar milho / (v iv sg) emparelhar, responder, discutir
ke - (n) sobra, resto / (v iv) futuro, fazer, dizer
ker - (o) cuidado / (n) presente de comida, cama de barbante de urtiga
kre - (n) balaio, coxa / (v tv pl) cortar, ceifar
kren - (mod) quase / (v iv sg) escapar / (v tv sg) perder
krenkren - (v iv pl) escapar / (v tv pl) perder
kur - (o) ligeiro, rapido! / (n) roupa, vestido
mé - (mod) fazer diariamente, ligeiro, gostar de fazer / (n) carneiro
mẽ - (mod) muito, ligeiro / (v tv sg) cheirar, escutar, sentir, tocar
mro - (v iv sg) tomar banho / (v tv sg) por de molho
mũ - (a) fazendo, narrativo, acao unica, consequencia / (v iv pl) andar
mỹ - (cir) para / (su) pergunta
mỹr - (o) verdadeiro, e certo / (conj) ?
ne - (su) originador / (v tv sg) enterrar / (n) enterrado / de / (int) o que?
nẽ - (a) sera que? / (v tv) cobrir encostado
nig - (n) lagoa, poco / (v tv sg) dar pontape
nĩ - (a) na situacao de, sentado, ter, ter a obrigacao, condicao / (n) carne / (v iv sg) sentar
nỹ - (a) deitado / (su) topico na pergunta / (n) mae, irma da mae / (v iv sg) deitar-se
nỹtĩ - (a) sendo, ter / (v iv pl) existir
pẽ - (mod) muito, o verdadeiro / (n) braco
ra - (cir) para, na direcao de, apesar de / (n) queixo / (a) faca ja!
rã - (cir) perto, por baixo de / (n) sol, maduro / (v iv sg) entrar, comecar, estar perto de
rán - (n) encosta, declive, barranco, subida, perau / (v tv sg) escrever
rãké - (v iv) chegando tarde / tarde ?
rãnhrãj - (n) trabalho / (v tv pl) trabalhar
re - (n) grama, campo, fora de casa / (v iv) encher-se de bebida / (v iv pl) descer
sa - (a) pendurado / (n) sal, cachoeira / (v iv) pendurado / (v tv) pendurar, pregar
se - (n) quati / (v tv sg) amarrar, prender / (ij) puxa!
tãnh - (n) morador, dono / (v iv) fazer forte, habituar-se
ti - (pr pes) ele / (pr dem) o
tógfĩn - (n) gaviao / (v tv sg) atar, amarrar
tỹ - (cir) com, por / (erg) / (ex) / (top)
vãn - (n) taquara / (v tv sg) carregar coisa comprida
vãnh - (mod) nao querer ou poder fazer algo / (n) capoeira, mato, plantas curtas
ve - (n) aparencia, natureza, irma de um homem, primeiro / (v iv sg) parecer ser / (v tv sg) ver, enxergar
vo - (v iv) barulho de andar no mato, coracao batendo / (n) macuco
vỹ - (su) topico / (o) sera que e? / (voc) homem! moco!

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T03:59:39Z

Drosset1: /* What we did */

==The project: RBMT or NMT?==
===Background===
Kaingang is a language spoken in the southern Brazil region. It has roughly 20,000 speakers, many of whom are bilingual with Portuguese. Kaingang is one the most well-documented languages of indigenous peoples of Brazil; however, in our project, we were only able to use a [https://www.sil.org/resources/archives/42876 Kaingang-Portuguese dictionary] and a [https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf Kaingang new testament bible] (the other resources were written in languages we didn't read).
===Idea===
We wanted to find out how well a neural machine translation could work with a language that didn't have millions of parallel lines (in fact, we have ~10,000). We also wanted to compare it to our rule-based transducer, which we've been making additions to since the end of the semester.

===Motivation===
This is the only (at least easily accessible when searching) machine translation available between Kaingang and another language. The motivation behind our project was to possibly allow the Kaingang people to not have to choose between teaching Kaingang or Portuguese to later generations. We hoped (if at some point one of the translators was state-of-the-art) any need for Portuguese could be accessed through the translator, which would be an incentive to keep the language Kaingang alive.

==What we did==
* Added many more features (such as much larger coverage, many more words added to the bilingual dictionary, new disambiguation rules, and 2 new structural transfer rules) and improved performance of rule-based transducer
* Trained a neural network with 75% of our parallel corpus and tested with 25% using OpenNMT
* Documented ambiguity in Kaingang
* Created parallel corpora for the whole new testament between Kaingang and Portuguese!

==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]

[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation Metric==
Our new evaluation accounts for the fact that words in the Apertium translation sentences has # and * symbols. The accuracy script looks something like this:
correct_word_percentage = []
for i in range(len(true_translations)):
correct_word_count = 0
for word in prediction[i].split():
if word.strip("\n").strip("*").strip("#") in true_translations[i]:
correct_word_count += 1
correct_word_percentage.append(correct_word_percentage/len(prediction[i].split()))
average_accuracy = sum(correct_word_percentage)/len(correct_word_percentage)
print(average_accuracy)
===Accuracy===
Initial RBMT accuracy (actually higher than poster because of new evaluation metric): 6.78%

Final RBMT accuracy: 9.91%

Initial NMT accuracy: 21.4%

Final NMT accuracy: (currently training on new larger corpus, so no final accuracy yet)

==Specs for rule-based==
===Coverage===

* Coverage over new testament:
$ aq-covtest ../ling073-kgp-corpus/kgp.corpus.large.txt kgp.automorf.bin
Number of tokenised words in the corpus: 406362
Coverage: 92.93%
Top unknown words in the corpus:
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
270 nĩnh
269 tũꞌ
Translation time: 1.422363042831421 seconds
* Coverage over dictionary sentences:
$ aq-covtest ../ling073-kgp-por-corpus/kgp.sentences.txt kgp.automorf.bin
Number of tokenised words in the corpus: 32659
Coverage: 92.72%
Top unknown words in the corpus:
30 nĩm
27 vég
26 isóg
24 jagy
23 prũ
23 tãg
22 kyr
20 mỹnh
20 régre
18 pir
18 féj
17 jã
17 Kusa
17 nĩgé
16 kãtĩ
16 jamã
15 venh
15 gãnh
15 jakré
14 nug
Translation time: 0.1878523826599121 seconds

===Word Error Rate===

==Further work==
===Documented ambiguity in Kaingang===
We haven't employed rules for the vast majority of these but we compiled a list of all ambiguous words in the translator!
ag - (pr pes) eles / (pr dem) os
e - (mod) muito / (v iv) fazer, causar
ẽmĩ - (n) pao de milho / (v iv) tatear, apalpar
fa - (n) perna, planta, amargo / (v tv sg) lavar roupa
fag - (pr pes) elas / (pr dem) as
fi - (pr pes) ela / (pr dem) a
ge - (mod) entao, tambem / (n) semelhante / (v iv) fazer igual
gé - (mod) tambem / (v tv pl) levar, carregar
gunhgunh - (n) picao preto / (v tv pl) fincar
han - (v iv sg) sarar, melhorar / (v tv sg) fazer
he - (ij) sim / (v tv) dizer
hur - (o) ja / (v iv sg) caimbra
hỹn - (o) de certo, provavelmente / (int) onde?
jãn - (v tv sg) rasgar / (v iv sg) rezar, cantar hinos religiosos
jãvãnh - (mod) recusar, nao saber fazer / (v tv sg) esperar, procurar
jé - (su) futuro / (n) reza, hino
jẽn - (v i) comer / (v tv) dar de comer
jo - (conj) entao / (cir) antes, na frente
jũn - (n) palmito / (v tv) embrabecer
kafãn - (v tv sg) descascar milho / (v iv sg) emparelhar, responder, discutir
ke - (n) sobra, resto / (v iv) futuro, fazer, dizer
ker - (o) cuidado / (n) presente de comida, cama de barbante de urtiga
kre - (n) balaio, coxa / (v tv pl) cortar, ceifar
kren - (mod) quase / (v iv sg) escapar / (v tv sg) perder
krenkren - (v iv pl) escapar / (v tv pl) perder
kur - (o) ligeiro, rapido! / (n) roupa, vestido
mé - (mod) fazer diariamente, ligeiro, gostar de fazer / (n) carneiro
mẽ - (mod) muito, ligeiro / (v tv sg) cheirar, escutar, sentir, tocar
mro - (v iv sg) tomar banho / (v tv sg) por de molho
mũ - (a) fazendo, narrativo, acao unica, consequencia / (v iv pl) andar
mỹ - (cir) para / (su) pergunta
mỹr - (o) verdadeiro, e certo / (conj) ?
ne - (su) originador / (v tv sg) enterrar / (n) enterrado / de / (int) o que?
nẽ - (a) sera que? / (v tv) cobrir encostado
nig - (n) lagoa, poco / (v tv sg) dar pontape
nĩ - (a) na situacao de, sentado, ter, ter a obrigacao, condicao / (n) carne / (v iv sg) sentar
nỹ - (a) deitado / (su) topico na pergunta / (n) mae, irma da mae / (v iv sg) deitar-se
nỹtĩ - (a) sendo, ter / (v iv pl) existir
pẽ - (mod) muito, o verdadeiro / (n) braco
ra - (cir) para, na direcao de, apesar de / (n) queixo / (a) faca ja!
rã - (cir) perto, por baixo de / (n) sol, maduro / (v iv sg) entrar, comecar, estar perto de
rán - (n) encosta, declive, barranco, subida, perau / (v tv sg) escrever
rãké - (v iv) chegando tarde / tarde ?
rãnhrãj - (n) trabalho / (v tv pl) trabalhar
re - (n) grama, campo, fora de casa / (v iv) encher-se de bebida / (v iv pl) descer
sa - (a) pendurado / (n) sal, cachoeira / (v iv) pendurado / (v tv) pendurar, pregar
se - (n) quati / (v tv sg) amarrar, prender / (ij) puxa!
tãnh - (n) morador, dono / (v iv) fazer forte, habituar-se
ti - (pr pes) ele / (pr dem) o
tógfĩn - (n) gaviao / (v tv sg) atar, amarrar
tỹ - (cir) com, por / (erg) / (ex) / (top)
vãn - (n) taquara / (v tv sg) carregar coisa comprida
vãnh - (mod) nao querer ou poder fazer algo / (n) capoeira, mato, plantas curtas
ve - (n) aparencia, natureza, irma de um homem, primeiro / (v iv sg) parecer ser / (v tv sg) ver, enxergar
vo - (v iv) barulho de andar no mato, coracao batendo / (n) macuco
vỹ - (su) topico / (o) sera que e? / (voc) homem! moco!

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T03:56:44Z

Drosset1: /* Documented ambiguity in Kaingang */

==The project: RBMT or NMT?==
===Background===
Kaingang is a language spoken in the southern Brazil region. It has roughly 20,000 speakers, many of whom are bilingual with Portuguese. Kaingang is one the most well-documented languages of indigenous peoples of Brazil; however, in our project, we were only able to use a [https://www.sil.org/resources/archives/42876 Kaingang-Portuguese dictionary] and a [https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf Kaingang new testament bible] (the other resources were written in languages we didn't read).
===Idea===
We wanted to find out how well a neural machine translation could work with a language that didn't have millions of parallel lines (in fact, we have ~10,000). We also wanted to compare it to our rule-based transducer, which we've been making additions to since the end of the semester.

===Motivation===
This is the only (at least easily accessible when searching) machine translation available between Kaingang and another language. The motivation behind our project was to possibly allow the Kaingang people to not have to choose between teaching Kaingang or Portuguese to later generations. We hoped (if at some point one of the translators was state-of-the-art) any need for Portuguese could be accessed through the translator, which would be an incentive to keep the language Kaingang alive.

==What we did==
* Added many more features (such as much larger coverage, new disambiguation rules, and 2 new structural transfer rules) and improved performance of rule-based transducer
* Trained a neural network with 75% of our parallel corpus and tested with 25% using OpenNMT
* Documented ambiguity in Kaingang
* Created parallel corpora for the whole new testament between Kaingang and Portuguese!

==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]

[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation Metric==
Our new evaluation accounts for the fact that words in the Apertium translation sentences has # and * symbols. The accuracy script looks something like this:
correct_word_percentage = []
for i in range(len(true_translations)):
correct_word_count = 0
for word in prediction[i].split():
if word.strip("\n").strip("*").strip("#") in true_translations[i]:
correct_word_count += 1
correct_word_percentage.append(correct_word_percentage/len(prediction[i].split()))
average_accuracy = sum(correct_word_percentage)/len(correct_word_percentage)
print(average_accuracy)
===Accuracy===
Initial RBMT accuracy (actually higher than poster because of new evaluation metric): 6.78%

Final RBMT accuracy: 9.91%

Initial NMT accuracy: 21.4%

Final NMT accuracy: (currently training on new larger corpus, so no final accuracy yet)

==Specs for rule-based==
===Coverage===

* Coverage over new testament:
$ aq-covtest ../ling073-kgp-corpus/kgp.corpus.large.txt kgp.automorf.bin
Number of tokenised words in the corpus: 406362
Coverage: 92.93%
Top unknown words in the corpus:
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
270 nĩnh
269 tũꞌ
Translation time: 1.422363042831421 seconds
* Coverage over dictionary sentences:
$ aq-covtest ../ling073-kgp-por-corpus/kgp.sentences.txt kgp.automorf.bin
Number of tokenised words in the corpus: 32659
Coverage: 92.72%
Top unknown words in the corpus:
30 nĩm
27 vég
26 isóg
24 jagy
23 prũ
23 tãg
22 kyr
20 mỹnh
20 régre
18 pir
18 féj
17 jã
17 Kusa
17 nĩgé
16 kãtĩ
16 jamã
15 venh
15 gãnh
15 jakré
14 nug
Translation time: 0.1878523826599121 seconds

===Word Error Rate===

==Further work==
===Documented ambiguity in Kaingang===
We haven't employed rules for the vast majority of these but we compiled a list of all ambiguous words in the translator!
ag - (pr pes) eles / (pr dem) os
e - (mod) muito / (v iv) fazer, causar
ẽmĩ - (n) pao de milho / (v iv) tatear, apalpar
fa - (n) perna, planta, amargo / (v tv sg) lavar roupa
fag - (pr pes) elas / (pr dem) as
fi - (pr pes) ela / (pr dem) a
ge - (mod) entao, tambem / (n) semelhante / (v iv) fazer igual
gé - (mod) tambem / (v tv pl) levar, carregar
gunhgunh - (n) picao preto / (v tv pl) fincar
han - (v iv sg) sarar, melhorar / (v tv sg) fazer
he - (ij) sim / (v tv) dizer
hur - (o) ja / (v iv sg) caimbra
hỹn - (o) de certo, provavelmente / (int) onde?
jãn - (v tv sg) rasgar / (v iv sg) rezar, cantar hinos religiosos
jãvãnh - (mod) recusar, nao saber fazer / (v tv sg) esperar, procurar
jé - (su) futuro / (n) reza, hino
jẽn - (v i) comer / (v tv) dar de comer
jo - (conj) entao / (cir) antes, na frente
jũn - (n) palmito / (v tv) embrabecer
kafãn - (v tv sg) descascar milho / (v iv sg) emparelhar, responder, discutir
ke - (n) sobra, resto / (v iv) futuro, fazer, dizer
ker - (o) cuidado / (n) presente de comida, cama de barbante de urtiga
kre - (n) balaio, coxa / (v tv pl) cortar, ceifar
kren - (mod) quase / (v iv sg) escapar / (v tv sg) perder
krenkren - (v iv pl) escapar / (v tv pl) perder
kur - (o) ligeiro, rapido! / (n) roupa, vestido
mé - (mod) fazer diariamente, ligeiro, gostar de fazer / (n) carneiro
mẽ - (mod) muito, ligeiro / (v tv sg) cheirar, escutar, sentir, tocar
mro - (v iv sg) tomar banho / (v tv sg) por de molho
mũ - (a) fazendo, narrativo, acao unica, consequencia / (v iv pl) andar
mỹ - (cir) para / (su) pergunta
mỹr - (o) verdadeiro, e certo / (conj) ?
ne - (su) originador / (v tv sg) enterrar / (n) enterrado / de / (int) o que?
nẽ - (a) sera que? / (v tv) cobrir encostado
nig - (n) lagoa, poco / (v tv sg) dar pontape
nĩ - (a) na situacao de, sentado, ter, ter a obrigacao, condicao / (n) carne / (v iv sg) sentar
nỹ - (a) deitado / (su) topico na pergunta / (n) mae, irma da mae / (v iv sg) deitar-se
nỹtĩ - (a) sendo, ter / (v iv pl) existir
pẽ - (mod) muito, o verdadeiro / (n) braco
ra - (cir) para, na direcao de, apesar de / (n) queixo / (a) faca ja!
rã - (cir) perto, por baixo de / (n) sol, maduro / (v iv sg) entrar, comecar, estar perto de
rán - (n) encosta, declive, barranco, subida, perau / (v tv sg) escrever
rãké - (v iv) chegando tarde / tarde ?
rãnhrãj - (n) trabalho / (v tv pl) trabalhar
re - (n) grama, campo, fora de casa / (v iv) encher-se de bebida / (v iv pl) descer
sa - (a) pendurado / (n) sal, cachoeira / (v iv) pendurado / (v tv) pendurar, pregar
se - (n) quati / (v tv sg) amarrar, prender / (ij) puxa!
tãnh - (n) morador, dono / (v iv) fazer forte, habituar-se
ti - (pr pes) ele / (pr dem) o
tógfĩn - (n) gaviao / (v tv sg) atar, amarrar
tỹ - (cir) com, por / (erg) / (ex) / (top)
vãn - (n) taquara / (v tv sg) carregar coisa comprida
vãnh - (mod) nao querer ou poder fazer algo / (n) capoeira, mato, plantas curtas
ve - (n) aparencia, natureza, irma de um homem, primeiro / (v iv sg) parecer ser / (v tv sg) ver, enxergar
vo - (v iv) barulho de andar no mato, coracao batendo / (n) macuco
vỹ - (su) topico / (o) sera que e? / (voc) homem! moco!

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T03:53:06Z

Drosset1: /* Documented ambiguity in Kaingang */

==The project: RBMT or NMT?==
===Background===
Kaingang is a language spoken in the southern Brazil region. It has roughly 20,000 speakers, many of whom are bilingual with Portuguese. Kaingang is one the most well-documented languages of indigenous peoples of Brazil; however, in our project, we were only able to use a [https://www.sil.org/resources/archives/42876 Kaingang-Portuguese dictionary] and a [https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf Kaingang new testament bible] (the other resources were written in languages we didn't read).
===Idea===
We wanted to find out how well a neural machine translation could work with a language that didn't have millions of parallel lines (in fact, we have ~10,000). We also wanted to compare it to our rule-based transducer, which we've been making additions to since the end of the semester.

===Motivation===
This is the only (at least easily accessible when searching) machine translation available between Kaingang and another language. The motivation behind our project was to possibly allow the Kaingang people to not have to choose between teaching Kaingang or Portuguese to later generations. We hoped (if at some point one of the translators was state-of-the-art) any need for Portuguese could be accessed through the translator, which would be an incentive to keep the language Kaingang alive.

==What we did==
* Added many more features (such as much larger coverage, new disambiguation rules, and 2 new structural transfer rules) and improved performance of rule-based transducer
* Trained a neural network with 75% of our parallel corpus and tested with 25% using OpenNMT
* Documented ambiguity in Kaingang
* Created parallel corpora for the whole new testament between Kaingang and Portuguese!

==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]

[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation Metric==
Our new evaluation accounts for the fact that words in the Apertium translation sentences has # and * symbols. The accuracy script looks something like this:
correct_word_percentage = []
for i in range(len(true_translations)):
correct_word_count = 0
for word in prediction[i].split():
if word.strip("\n").strip("*").strip("#") in true_translations[i]:
correct_word_count += 1
correct_word_percentage.append(correct_word_percentage/len(prediction[i].split()))
average_accuracy = sum(correct_word_percentage)/len(correct_word_percentage)
print(average_accuracy)
===Accuracy===
Initial RBMT accuracy (actually higher than poster because of new evaluation metric): 6.78%

Final RBMT accuracy: 9.91%

Initial NMT accuracy: 21.4%

Final NMT accuracy: (currently training on new larger corpus, so no final accuracy yet)

==Specs for rule-based==
===Coverage===

* Coverage over new testament:
$ aq-covtest ../ling073-kgp-corpus/kgp.corpus.large.txt kgp.automorf.bin
Number of tokenised words in the corpus: 406362
Coverage: 92.93%
Top unknown words in the corpus:
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
270 nĩnh
269 tũꞌ
Translation time: 1.422363042831421 seconds
* Coverage over dictionary sentences:
$ aq-covtest ../ling073-kgp-por-corpus/kgp.sentences.txt kgp.automorf.bin
Number of tokenised words in the corpus: 32659
Coverage: 92.72%
Top unknown words in the corpus:
30 nĩm
27 vég
26 isóg
24 jagy
23 prũ
23 tãg
22 kyr
20 mỹnh
20 régre
18 pir
18 féj
17 jã
17 Kusa
17 nĩgé
16 kãtĩ
16 jamã
15 venh
15 gãnh
15 jakré
14 nug
Translation time: 0.1878523826599121 seconds

===Word Error Rate===

==Further work==
===Documented ambiguity in Kaingang===
We haven't employed rules for the vast majority of these but we compiled a list of all ambiguous words in the translator!
ag - (pr pes) eles / (pr dem) os
ã -
ãjag -
e - (mod) muito / (v iv) fazer, causar
eg -
ẽmĩ - (n) pao de milho / (v iv) tatear, apalpar
fa - (n) perna, planta, amargo / (v tv sg) lavar roupa
fag - (pr pes) elas / (pr dem) as
fi - (pr pes) ela / (pr dem) a
ge - (mod) entao, tambem / (n) semelhante / (v iv) fazer igual
gé - (mod) tambem / (v tv pl) levar, carregar
gunhgunh - (n) picao preto / (v tv pl) fincar
han - (v iv sg) sarar, melhorar / (v tv sg) fazer
he - (ij) sim / (v tv) dizer
hur - (o) ja / (v iv sg) caimbra
hỹn - (o) de certo, provavelmente / (int) onde?
inh -
jãn -
jãvãnh - (mod) recusar, nao saber fazer / (v tv sg) esperar, procurar
jé - (su) futuro / (n) reza, hino
jẽn -
jo - (conj) entao / (cir) antes, na frente
jũn - (n) palmito / (v tv) embrabecer
kafãn -
ke - (n) sobra, resto / (v iv) futuro, fazer, dizer
ker - (o) cuidado / (n) presente de comida, cama de barbante de urtiga
kre - (n) balaio, coxa / (v tv pl) cortar, ceifar
kren - (mod) quase / (v iv sg) escapar / (v tv sg) perder
krenkren - (v iv pl) escapar / (v tv pl) perder
kur - (o) ligeiro, rapido! / (n) roupa, vestido
mé - (mod) fazer diariamente, ligeiro, gostar de fazer / (n) carneiro
mẽ - (mod) muito, ligeiro / (v tv sg) cheirar, escutar, sentir, tocar
mro - (v iv sg) tomar banho / (v tv sg) por de molho
mũ - (a) fazendo, narrativo, acao unica, consequencia / (v iv pl) andar
mỹ - (cir) para / (su) pergunta
mỹr - (o) verdadeiro, e certo / (conj) ?
ne - (su) originador / (v tv sg) enterrar / (n) enterrado / de / (int) o que?
nẽ - (a) sera que? / (v tv) cobrir encostado
nig - (n) lagoa, poco / (v tv sg) dar pontape
nĩ - (a) na situacao de, sentado, ter, ter a obrigacao, condicao / (n) carne / (v iv sg) sentar
nỹ - (a) deitado / (su) topico na pergunta / (n) mae, irma da mae / (v iv sg) deitar-se
nỹtĩ - (a) sendo, ter / (v iv pl) existir
pẽ - (mod) muito, o verdadeiro / (n) braco
ra - (cir) para, na direcao de, apesar de / (n) queixo / (a) faca ja!
rã - (cir) perto, por baixo de / (n) sol, maduro / (v iv sg) entrar, comecar, estar perto de
rán - (n) encosta, declive, barranco, subida, perau / (v tv sg) escrever
rãké - (v iv) chegando tarde / tarde ?
rãnhrãj - (n) trabalho / (v tv pl) trabalhar
re - (n) grama, campo, fora de casa / (v iv) encher-se de bebida / (v iv pl) descer
sa - (a) pendurado / (n) sal, cachoeira / (v iv) pendurado / (v tv) pendurar, pregar
se - (n) quati / (v tv sg) amarrar, prender / (ij) puxa!
tãnh - (n) morador, dono / (v iv) fazer forte, habituar-se
ti - (pr pes) ele / (pr dem) o
tógfĩn - (n) gaviao / (v tv sg) atar, amarrar
tỹ - (cir) com, por / (erg) / (ex) / (top)
vãn - (n) taquara / (v tv sg) carregar coisa comprida
vãnh - (mod) nao querer ou poder fazer algo / (n) capoeira, mato, plantas curtas
ve - (n) aparencia, natureza, irma de um homem, primeiro / (v iv sg) parecer ser / (v tv sg) ver, enxergar
vo - (v iv) barulho de andar no mato, coracao batendo / (n) macuco
vỹ - (su) topico / (o) sera que e? / (voc) homem! moco!

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T03:51:51Z

Drosset1: /* What we did */

==The project: RBMT or NMT?==
===Background===
Kaingang is a language spoken in the southern Brazil region. It has roughly 20,000 speakers, many of whom are bilingual with Portuguese. Kaingang is one the most well-documented languages of indigenous peoples of Brazil; however, in our project, we were only able to use a [https://www.sil.org/resources/archives/42876 Kaingang-Portuguese dictionary] and a [https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf Kaingang new testament bible] (the other resources were written in languages we didn't read).
===Idea===
We wanted to find out how well a neural machine translation could work with a language that didn't have millions of parallel lines (in fact, we have ~10,000). We also wanted to compare it to our rule-based transducer, which we've been making additions to since the end of the semester.

===Motivation===
This is the only (at least easily accessible when searching) machine translation available between Kaingang and another language. The motivation behind our project was to possibly allow the Kaingang people to not have to choose between teaching Kaingang or Portuguese to later generations. We hoped (if at some point one of the translators was state-of-the-art) any need for Portuguese could be accessed through the translator, which would be an incentive to keep the language Kaingang alive.

==What we did==
* Added many more features (such as much larger coverage, new disambiguation rules, and 2 new structural transfer rules) and improved performance of rule-based transducer
* Trained a neural network with 75% of our parallel corpus and tested with 25% using OpenNMT
* Documented ambiguity in Kaingang
* Created parallel corpora for the whole new testament between Kaingang and Portuguese!

==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]

[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation Metric==
Our new evaluation accounts for the fact that words in the Apertium translation sentences has # and * symbols. The accuracy script looks something like this:
correct_word_percentage = []
for i in range(len(true_translations)):
correct_word_count = 0
for word in prediction[i].split():
if word.strip("\n").strip("*").strip("#") in true_translations[i]:
correct_word_count += 1
correct_word_percentage.append(correct_word_percentage/len(prediction[i].split()))
average_accuracy = sum(correct_word_percentage)/len(correct_word_percentage)
print(average_accuracy)
===Accuracy===
Initial RBMT accuracy (actually higher than poster because of new evaluation metric): 6.78%

Final RBMT accuracy: 9.91%

Initial NMT accuracy: 21.4%

Final NMT accuracy: (currently training on new larger corpus, so no final accuracy yet)

==Specs for rule-based==
===Coverage===

* Coverage over new testament:
$ aq-covtest ../ling073-kgp-corpus/kgp.corpus.large.txt kgp.automorf.bin
Number of tokenised words in the corpus: 406362
Coverage: 92.93%
Top unknown words in the corpus:
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
270 nĩnh
269 tũꞌ
Translation time: 1.422363042831421 seconds
* Coverage over dictionary sentences:
$ aq-covtest ../ling073-kgp-por-corpus/kgp.sentences.txt kgp.automorf.bin
Number of tokenised words in the corpus: 32659
Coverage: 92.72%
Top unknown words in the corpus:
30 nĩm
27 vég
26 isóg
24 jagy
23 prũ
23 tãg
22 kyr
20 mỹnh
20 régre
18 pir
18 féj
17 jã
17 Kusa
17 nĩgé
16 kãtĩ
16 jamã
15 venh
15 gãnh
15 jakré
14 nug
Translation time: 0.1878523826599121 seconds

===Word Error Rate===

==Documented ambiguity in Kaingang==
ag - (pr pes) eles / (pr dem) os
ã -
ãjag -
e - (mod) muito / (v iv) fazer, causar
eg -
ẽmĩ - (n) pao de milho / (v iv) tatear, apalpar
fa - (n) perna, planta, amargo / (v tv sg) lavar roupa
fag - (pr pes) elas / (pr dem) as
fi - (pr pes) ela / (pr dem) a
ge - (mod) entao, tambem / (n) semelhante / (v iv) fazer igual
gé - (mod) tambem / (v tv pl) levar, carregar
gunhgunh - (n) picao preto / (v tv pl) fincar
han - (v iv sg) sarar, melhorar / (v tv sg) fazer
he - (ij) sim / (v tv) dizer
hur - (o) ja / (v iv sg) caimbra
hỹn - (o) de certo, provavelmente / (int) onde?
inh -
jãn -
jãvãnh - (mod) recusar, nao saber fazer / (v tv sg) esperar, procurar
jé - (su) futuro / (n) reza, hino
jẽn -
jo - (conj) entao / (cir) antes, na frente
jũn - (n) palmito / (v tv) embrabecer
kafãn -
ke - (n) sobra, resto / (v iv) futuro, fazer, dizer
ker - (o) cuidado / (n) presente de comida, cama de barbante de urtiga
kre - (n) balaio, coxa / (v tv pl) cortar, ceifar
kren - (mod) quase / (v iv sg) escapar / (v tv sg) perder
krenkren - (v iv pl) escapar / (v tv pl) perder
kur - (o) ligeiro, rapido! / (n) roupa, vestido
mé - (mod) fazer diariamente, ligeiro, gostar de fazer / (n) carneiro
mẽ - (mod) muito, ligeiro / (v tv sg) cheirar, escutar, sentir, tocar
mro - (v iv sg) tomar banho / (v tv sg) por de molho
mũ - (a) fazendo, narrativo, acao unica, consequencia / (v iv pl) andar
mỹ - (cir) para / (su) pergunta
mỹr - (o) verdadeiro, e certo / (conj) ?
ne - (su) originador / (v tv sg) enterrar / (n) enterrado / de / (int) o que?
nẽ - (a) sera que? / (v tv) cobrir encostado
nig - (n) lagoa, poco / (v tv sg) dar pontape
nĩ - (a) na situacao de, sentado, ter, ter a obrigacao, condicao / (n) carne / (v iv sg) sentar
nỹ - (a) deitado / (su) topico na pergunta / (n) mae, irma da mae / (v iv sg) deitar-se
nỹtĩ - (a) sendo, ter / (v iv pl) existir
pẽ - (mod) muito, o verdadeiro / (n) braco
ra - (cir) para, na direcao de, apesar de / (n) queixo / (a) faca ja!
rã - (cir) perto, por baixo de / (n) sol, maduro / (v iv sg) entrar, comecar, estar perto de
rán - (n) encosta, declive, barranco, subida, perau / (v tv sg) escrever
rãké - (v iv) chegando tarde / tarde ?
rãnhrãj - (n) trabalho / (v tv pl) trabalhar
re - (n) grama, campo, fora de casa / (v iv) encher-se de bebida / (v iv pl) descer
sa - (a) pendurado / (n) sal, cachoeira / (v iv) pendurado / (v tv) pendurar, pregar
se - (n) quati / (v tv sg) amarrar, prender / (ij) puxa!
tãnh - (n) morador, dono / (v iv) fazer forte, habituar-se
ti - (pr pes) ele / (pr dem) o
tógfĩn - (n) gaviao / (v tv sg) atar, amarrar
tỹ - (cir) com, por / (erg) / (ex) / (top)
vãn - (n) taquara / (v tv sg) carregar coisa comprida
vãnh - (mod) nao querer ou poder fazer algo / (n) capoeira, mato, plantas curtas
ve - (n) aparencia, natureza, irma de um homem, primeiro / (v iv sg) parecer ser / (v tv sg) ver, enxergar
vo - (v iv) barulho de andar no mato, coracao batendo / (n) macuco
vỹ - (su) topico / (o) sera que e? / (voc) homem! moco!

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T03:49:21Z

Drosset1: /* Accuracy */

==The project: RBMT or NMT?==
===Background===
Kaingang is a language spoken in the southern Brazil region. It has roughly 20,000 speakers, many of whom are bilingual with Portuguese. Kaingang is one the most well-documented languages of indigenous peoples of Brazil; however, in our project, we were only able to use a [https://www.sil.org/resources/archives/42876 Kaingang-Portuguese dictionary] and a [https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf Kaingang new testament bible] (the other resources were written in languages we didn't read).
===Idea===
We wanted to find out how well a neural machine translation could work with a language that didn't have millions of parallel lines (in fact, we have ~10,000). We also wanted to compare it to our rule-based transducer, which we've been making additions to since the end of the semester.

===Motivation===
This is the only (at least easily accessible when searching) machine translation available between Kaingang and another language. The motivation behind our project was to possibly allow the Kaingang people to not have to choose between teaching Kaingang or Portuguese to later generations. We hoped (if at some point one of the translators was state-of-the-art) any need for Portuguese could be accessed through the translator, which would be an incentive to keep the language Kaingang alive.

==What we did==
* Added many more features to and improved performance of rule-based transducer
* Trained a neural network with 75% of our parallel corpus and tested with 25% using OpenNMT
* Documented ambiguity in Kaingang
* Created parallel corpora for the whole new testament between Kaingang and Portuguese!
==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]

[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation Metric==
Our new evaluation accounts for the fact that words in the Apertium translation sentences has # and * symbols. The accuracy script looks something like this:
correct_word_percentage = []
for i in range(len(true_translations)):
correct_word_count = 0
for word in prediction[i].split():
if word.strip("\n").strip("*").strip("#") in true_translations[i]:
correct_word_count += 1
correct_word_percentage.append(correct_word_percentage/len(prediction[i].split()))
average_accuracy = sum(correct_word_percentage)/len(correct_word_percentage)
print(average_accuracy)
===Accuracy===
Initial RBMT accuracy (actually higher than poster because of new evaluation metric): 6.78%

Final RBMT accuracy: 9.91%

Initial NMT accuracy: 21.4%

Final NMT accuracy: (currently training on new larger corpus, so no final accuracy yet)

==Specs for rule-based==
===Coverage===

* Coverage over new testament:
$ aq-covtest ../ling073-kgp-corpus/kgp.corpus.large.txt kgp.automorf.bin
Number of tokenised words in the corpus: 406362
Coverage: 92.93%
Top unknown words in the corpus:
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
270 nĩnh
269 tũꞌ
Translation time: 1.422363042831421 seconds
* Coverage over dictionary sentences:
$ aq-covtest ../ling073-kgp-por-corpus/kgp.sentences.txt kgp.automorf.bin
Number of tokenised words in the corpus: 32659
Coverage: 92.72%
Top unknown words in the corpus:
30 nĩm
27 vég
26 isóg
24 jagy
23 prũ
23 tãg
22 kyr
20 mỹnh
20 régre
18 pir
18 féj
17 jã
17 Kusa
17 nĩgé
16 kãtĩ
16 jamã
15 venh
15 gãnh
15 jakré
14 nug
Translation time: 0.1878523826599121 seconds

===Word Error Rate===

==Documented ambiguity in Kaingang==
ag - (pr pes) eles / (pr dem) os
ã -
ãjag -
e - (mod) muito / (v iv) fazer, causar
eg -
ẽmĩ - (n) pao de milho / (v iv) tatear, apalpar
fa - (n) perna, planta, amargo / (v tv sg) lavar roupa
fag - (pr pes) elas / (pr dem) as
fi - (pr pes) ela / (pr dem) a
ge - (mod) entao, tambem / (n) semelhante / (v iv) fazer igual
gé - (mod) tambem / (v tv pl) levar, carregar
gunhgunh - (n) picao preto / (v tv pl) fincar
han - (v iv sg) sarar, melhorar / (v tv sg) fazer
he - (ij) sim / (v tv) dizer
hur - (o) ja / (v iv sg) caimbra
hỹn - (o) de certo, provavelmente / (int) onde?
inh -
jãn -
jãvãnh - (mod) recusar, nao saber fazer / (v tv sg) esperar, procurar
jé - (su) futuro / (n) reza, hino
jẽn -
jo - (conj) entao / (cir) antes, na frente
jũn - (n) palmito / (v tv) embrabecer
kafãn -
ke - (n) sobra, resto / (v iv) futuro, fazer, dizer
ker - (o) cuidado / (n) presente de comida, cama de barbante de urtiga
kre - (n) balaio, coxa / (v tv pl) cortar, ceifar
kren - (mod) quase / (v iv sg) escapar / (v tv sg) perder
krenkren - (v iv pl) escapar / (v tv pl) perder
kur - (o) ligeiro, rapido! / (n) roupa, vestido
mé - (mod) fazer diariamente, ligeiro, gostar de fazer / (n) carneiro
mẽ - (mod) muito, ligeiro / (v tv sg) cheirar, escutar, sentir, tocar
mro - (v iv sg) tomar banho / (v tv sg) por de molho
mũ - (a) fazendo, narrativo, acao unica, consequencia / (v iv pl) andar
mỹ - (cir) para / (su) pergunta
mỹr - (o) verdadeiro, e certo / (conj) ?
ne - (su) originador / (v tv sg) enterrar / (n) enterrado / de / (int) o que?
nẽ - (a) sera que? / (v tv) cobrir encostado
nig - (n) lagoa, poco / (v tv sg) dar pontape
nĩ - (a) na situacao de, sentado, ter, ter a obrigacao, condicao / (n) carne / (v iv sg) sentar
nỹ - (a) deitado / (su) topico na pergunta / (n) mae, irma da mae / (v iv sg) deitar-se
nỹtĩ - (a) sendo, ter / (v iv pl) existir
pẽ - (mod) muito, o verdadeiro / (n) braco
ra - (cir) para, na direcao de, apesar de / (n) queixo / (a) faca ja!
rã - (cir) perto, por baixo de / (n) sol, maduro / (v iv sg) entrar, comecar, estar perto de
rán - (n) encosta, declive, barranco, subida, perau / (v tv sg) escrever
rãké - (v iv) chegando tarde / tarde ?
rãnhrãj - (n) trabalho / (v tv pl) trabalhar
re - (n) grama, campo, fora de casa / (v iv) encher-se de bebida / (v iv pl) descer
sa - (a) pendurado / (n) sal, cachoeira / (v iv) pendurado / (v tv) pendurar, pregar
se - (n) quati / (v tv sg) amarrar, prender / (ij) puxa!
tãnh - (n) morador, dono / (v iv) fazer forte, habituar-se
ti - (pr pes) ele / (pr dem) o
tógfĩn - (n) gaviao / (v tv sg) atar, amarrar
tỹ - (cir) com, por / (erg) / (ex) / (top)
vãn - (n) taquara / (v tv sg) carregar coisa comprida
vãnh - (mod) nao querer ou poder fazer algo / (n) capoeira, mato, plantas curtas
ve - (n) aparencia, natureza, irma de um homem, primeiro / (v iv sg) parecer ser / (v tv sg) ver, enxergar
vo - (v iv) barulho de andar no mato, coracao batendo / (n) macuco
vỹ - (su) topico / (o) sera que e? / (voc) homem! moco!

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T03:48:44Z

Drosset1: /* Specs for rule-based */

==The project: RBMT or NMT?==
===Background===
Kaingang is a language spoken in the southern Brazil region. It has roughly 20,000 speakers, many of whom are bilingual with Portuguese. Kaingang is one the most well-documented languages of indigenous peoples of Brazil; however, in our project, we were only able to use a [https://www.sil.org/resources/archives/42876 Kaingang-Portuguese dictionary] and a [https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf Kaingang new testament bible] (the other resources were written in languages we didn't read).
===Idea===
We wanted to find out how well a neural machine translation could work with a language that didn't have millions of parallel lines (in fact, we have ~10,000). We also wanted to compare it to our rule-based transducer, which we've been making additions to since the end of the semester.

===Motivation===
This is the only (at least easily accessible when searching) machine translation available between Kaingang and another language. The motivation behind our project was to possibly allow the Kaingang people to not have to choose between teaching Kaingang or Portuguese to later generations. We hoped (if at some point one of the translators was state-of-the-art) any need for Portuguese could be accessed through the translator, which would be an incentive to keep the language Kaingang alive.

==What we did==
* Added many more features to and improved performance of rule-based transducer
* Trained a neural network with 75% of our parallel corpus and tested with 25% using OpenNMT
* Documented ambiguity in Kaingang
* Created parallel corpora for the whole new testament between Kaingang and Portuguese!
==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]

[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation Metric==
Our new evaluation accounts for the fact that words in the Apertium translation sentences has # and * symbols. The accuracy script looks something like this:
correct_word_percentage = []
for i in range(len(true_translations)):
correct_word_count = 0
for word in prediction[i].split():
if word.strip("\n").strip("*").strip("#") in true_translations[i]:
correct_word_count += 1
correct_word_percentage.append(correct_word_percentage/len(prediction[i].split()))
average_accuracy = sum(correct_word_percentage)/len(correct_word_percentage)
print(average_accuracy)
===Accuracy===
Initial RBMT accuracy:

Final RBMT accuracy: 9.91%

Initial NMT accuracy: 21.4%

Final NMT accuracy: (currently training on new larger corpus, so no final accuracy yet)

==Specs for rule-based==
===Coverage===

* Coverage over new testament:
$ aq-covtest ../ling073-kgp-corpus/kgp.corpus.large.txt kgp.automorf.bin
Number of tokenised words in the corpus: 406362
Coverage: 92.93%
Top unknown words in the corpus:
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
270 nĩnh
269 tũꞌ
Translation time: 1.422363042831421 seconds
* Coverage over dictionary sentences:
$ aq-covtest ../ling073-kgp-por-corpus/kgp.sentences.txt kgp.automorf.bin
Number of tokenised words in the corpus: 32659
Coverage: 92.72%
Top unknown words in the corpus:
30 nĩm
27 vég
26 isóg
24 jagy
23 prũ
23 tãg
22 kyr
20 mỹnh
20 régre
18 pir
18 féj
17 jã
17 Kusa
17 nĩgé
16 kãtĩ
16 jamã
15 venh
15 gãnh
15 jakré
14 nug
Translation time: 0.1878523826599121 seconds

===Word Error Rate===

==Documented ambiguity in Kaingang==
ag - (pr pes) eles / (pr dem) os
ã -
ãjag -
e - (mod) muito / (v iv) fazer, causar
eg -
ẽmĩ - (n) pao de milho / (v iv) tatear, apalpar
fa - (n) perna, planta, amargo / (v tv sg) lavar roupa
fag - (pr pes) elas / (pr dem) as
fi - (pr pes) ela / (pr dem) a
ge - (mod) entao, tambem / (n) semelhante / (v iv) fazer igual
gé - (mod) tambem / (v tv pl) levar, carregar
gunhgunh - (n) picao preto / (v tv pl) fincar
han - (v iv sg) sarar, melhorar / (v tv sg) fazer
he - (ij) sim / (v tv) dizer
hur - (o) ja / (v iv sg) caimbra
hỹn - (o) de certo, provavelmente / (int) onde?
inh -
jãn -
jãvãnh - (mod) recusar, nao saber fazer / (v tv sg) esperar, procurar
jé - (su) futuro / (n) reza, hino
jẽn -
jo - (conj) entao / (cir) antes, na frente
jũn - (n) palmito / (v tv) embrabecer
kafãn -
ke - (n) sobra, resto / (v iv) futuro, fazer, dizer
ker - (o) cuidado / (n) presente de comida, cama de barbante de urtiga
kre - (n) balaio, coxa / (v tv pl) cortar, ceifar
kren - (mod) quase / (v iv sg) escapar / (v tv sg) perder
krenkren - (v iv pl) escapar / (v tv pl) perder
kur - (o) ligeiro, rapido! / (n) roupa, vestido
mé - (mod) fazer diariamente, ligeiro, gostar de fazer / (n) carneiro
mẽ - (mod) muito, ligeiro / (v tv sg) cheirar, escutar, sentir, tocar
mro - (v iv sg) tomar banho / (v tv sg) por de molho
mũ - (a) fazendo, narrativo, acao unica, consequencia / (v iv pl) andar
mỹ - (cir) para / (su) pergunta
mỹr - (o) verdadeiro, e certo / (conj) ?
ne - (su) originador / (v tv sg) enterrar / (n) enterrado / de / (int) o que?
nẽ - (a) sera que? / (v tv) cobrir encostado
nig - (n) lagoa, poco / (v tv sg) dar pontape
nĩ - (a) na situacao de, sentado, ter, ter a obrigacao, condicao / (n) carne / (v iv sg) sentar
nỹ - (a) deitado / (su) topico na pergunta / (n) mae, irma da mae / (v iv sg) deitar-se
nỹtĩ - (a) sendo, ter / (v iv pl) existir
pẽ - (mod) muito, o verdadeiro / (n) braco
ra - (cir) para, na direcao de, apesar de / (n) queixo / (a) faca ja!
rã - (cir) perto, por baixo de / (n) sol, maduro / (v iv sg) entrar, comecar, estar perto de
rán - (n) encosta, declive, barranco, subida, perau / (v tv sg) escrever
rãké - (v iv) chegando tarde / tarde ?
rãnhrãj - (n) trabalho / (v tv pl) trabalhar
re - (n) grama, campo, fora de casa / (v iv) encher-se de bebida / (v iv pl) descer
sa - (a) pendurado / (n) sal, cachoeira / (v iv) pendurado / (v tv) pendurar, pregar
se - (n) quati / (v tv sg) amarrar, prender / (ij) puxa!
tãnh - (n) morador, dono / (v iv) fazer forte, habituar-se
ti - (pr pes) ele / (pr dem) o
tógfĩn - (n) gaviao / (v tv sg) atar, amarrar
tỹ - (cir) com, por / (erg) / (ex) / (top)
vãn - (n) taquara / (v tv sg) carregar coisa comprida
vãnh - (mod) nao querer ou poder fazer algo / (n) capoeira, mato, plantas curtas
ve - (n) aparencia, natureza, irma de um homem, primeiro / (v iv sg) parecer ser / (v tv sg) ver, enxergar
vo - (v iv) barulho de andar no mato, coracao batendo / (n) macuco
vỹ - (su) topico / (o) sera que e? / (voc) homem! moco!

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T03:42:08Z

Drosset1: /* Number of... */

==The project: RBMT or NMT?==
===Background===
Kaingang is a language spoken in the southern Brazil region. It has roughly 20,000 speakers, many of whom are bilingual with Portuguese. Kaingang is one the most well-documented languages of indigenous peoples of Brazil; however, in our project, we were only able to use a [https://www.sil.org/resources/archives/42876 Kaingang-Portuguese dictionary] and a [https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf Kaingang new testament bible] (the other resources were written in languages we didn't read).
===Idea===
We wanted to find out how well a neural machine translation could work with a language that didn't have millions of parallel lines (in fact, we have ~10,000). We also wanted to compare it to our rule-based transducer, which we've been making additions to since the end of the semester.

===Motivation===
This is the only (at least easily accessible when searching) machine translation available between Kaingang and another language. The motivation behind our project was to possibly allow the Kaingang people to not have to choose between teaching Kaingang or Portuguese to later generations. We hoped (if at some point one of the translators was state-of-the-art) any need for Portuguese could be accessed through the translator, which would be an incentive to keep the language Kaingang alive.

==What we did==
* Added many more features to and improved performance of rule-based transducer
* Trained a neural network with 75% of our parallel corpus and tested with 25% using OpenNMT
* Documented ambiguity in Kaingang
* Created parallel corpora for the whole new testament between Kaingang and Portuguese!
==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]

[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation Metric==
Our new evaluation accounts for the fact that words in the Apertium translation sentences has # and * symbols. The accuracy script looks something like this:
correct_word_percentage = []
for i in range(len(true_translations)):
correct_word_count = 0
for word in prediction[i].split():
if word.strip("\n").strip("*").strip("#") in true_translations[i]:
correct_word_count += 1
correct_word_percentage.append(correct_word_percentage/len(prediction[i].split()))
average_accuracy = sum(correct_word_percentage)/len(correct_word_percentage)
print(average_accuracy)
===Accuracy===
Initial RBMT accuracy:

Final RBMT accuracy: 9.91%

Initial NMT accuracy: 21.4%

Final NMT accuracy: (currently training on new larger corpus, so no final accuracy yet)

==Specs for rule-based==
===Coverage===
* Coverage over new testament:
$ aq-covtest ../ling073-kgp-corpus/kgp.corpus.large.txt kgp.automorf.bin
Number of tokenised words in the corpus: 406362
Coverage: 92.93%
Top unknown words in the corpus:
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
270 nĩnh
269 tũꞌ
Translation time: 1.422363042831421 seconds
* Coverage over dictionary sentences:
$ aq-covtest ../ling073-kgp-por-corpus/kgp.sentences.txt kgp.automorf.bin
Number of tokenised words in the corpus: 32659
Coverage: 92.72%
Top unknown words in the corpus:
30 nĩm
27 vég
26 isóg
24 jagy
23 prũ
23 tãg
22 kyr
20 mỹnh
20 régre
18 pir
18 féj
17 jã
17 Kusa
17 nĩgé
16 kãtĩ
16 jamã
15 venh
15 gãnh
15 jakré
14 nug
Translation time: 0.1878523826599121 seconds

===Word Error Rate===

==Documented ambiguity in Kaingang==
ag - (pr pes) eles / (pr dem) os
ã -
ãjag -
e - (mod) muito / (v iv) fazer, causar
eg -
ẽmĩ - (n) pao de milho / (v iv) tatear, apalpar
fa - (n) perna, planta, amargo / (v tv sg) lavar roupa
fag - (pr pes) elas / (pr dem) as
fi - (pr pes) ela / (pr dem) a
ge - (mod) entao, tambem / (n) semelhante / (v iv) fazer igual
gé - (mod) tambem / (v tv pl) levar, carregar
gunhgunh - (n) picao preto / (v tv pl) fincar
han - (v iv sg) sarar, melhorar / (v tv sg) fazer
he - (ij) sim / (v tv) dizer
hur - (o) ja / (v iv sg) caimbra
hỹn - (o) de certo, provavelmente / (int) onde?
inh -
jãn -
jãvãnh - (mod) recusar, nao saber fazer / (v tv sg) esperar, procurar
jé - (su) futuro / (n) reza, hino
jẽn -
jo - (conj) entao / (cir) antes, na frente
jũn - (n) palmito / (v tv) embrabecer
kafãn -
ke - (n) sobra, resto / (v iv) futuro, fazer, dizer
ker - (o) cuidado / (n) presente de comida, cama de barbante de urtiga
kre - (n) balaio, coxa / (v tv pl) cortar, ceifar
kren - (mod) quase / (v iv sg) escapar / (v tv sg) perder
krenkren - (v iv pl) escapar / (v tv pl) perder
kur - (o) ligeiro, rapido! / (n) roupa, vestido
mé - (mod) fazer diariamente, ligeiro, gostar de fazer / (n) carneiro
mẽ - (mod) muito, ligeiro / (v tv sg) cheirar, escutar, sentir, tocar
mro - (v iv sg) tomar banho / (v tv sg) por de molho
mũ - (a) fazendo, narrativo, acao unica, consequencia / (v iv pl) andar
mỹ - (cir) para / (su) pergunta
mỹr - (o) verdadeiro, e certo / (conj) ?
ne - (su) originador / (v tv sg) enterrar / (n) enterrado / de / (int) o que?
nẽ - (a) sera que? / (v tv) cobrir encostado
nig - (n) lagoa, poco / (v tv sg) dar pontape
nĩ - (a) na situacao de, sentado, ter, ter a obrigacao, condicao / (n) carne / (v iv sg) sentar
nỹ - (a) deitado / (su) topico na pergunta / (n) mae, irma da mae / (v iv sg) deitar-se
nỹtĩ - (a) sendo, ter / (v iv pl) existir
pẽ - (mod) muito, o verdadeiro / (n) braco
ra - (cir) para, na direcao de, apesar de / (n) queixo / (a) faca ja!
rã - (cir) perto, por baixo de / (n) sol, maduro / (v iv sg) entrar, comecar, estar perto de
rán - (n) encosta, declive, barranco, subida, perau / (v tv sg) escrever
rãké - (v iv) chegando tarde / tarde ?
rãnhrãj - (n) trabalho / (v tv pl) trabalhar
re - (n) grama, campo, fora de casa / (v iv) encher-se de bebida / (v iv pl) descer
sa - (a) pendurado / (n) sal, cachoeira / (v iv) pendurado / (v tv) pendurar, pregar
se - (n) quati / (v tv sg) amarrar, prender / (ij) puxa!
tãnh - (n) morador, dono / (v iv) fazer forte, habituar-se
ti - (pr pes) ele / (pr dem) o
tógfĩn - (n) gaviao / (v tv sg) atar, amarrar
tỹ - (cir) com, por / (erg) / (ex) / (top)
vãn - (n) taquara / (v tv sg) carregar coisa comprida
vãnh - (mod) nao querer ou poder fazer algo / (n) capoeira, mato, plantas curtas
ve - (n) aparencia, natureza, irma de um homem, primeiro / (v iv sg) parecer ser / (v tv sg) ver, enxergar
vo - (v iv) barulho de andar no mato, coracao batendo / (n) macuco
vỹ - (su) topico / (o) sera que e? / (voc) homem! moco!

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T03:33:12Z

Drosset1: /* Documented ambiguity in Kaingang */

==The project: RBMT or NMT?==
===Background===
Kaingang is a language spoken in the southern Brazil region. It has roughly 20,000 speakers, many of whom are bilingual with Portuguese. Kaingang is one the most well-documented languages of indigenous peoples of Brazil; however, in our project, we were only able to use a [https://www.sil.org/resources/archives/42876 Kaingang-Portuguese dictionary] and a [https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf Kaingang new testament bible] (the other resources were written in languages we didn't read).
===Idea===
We wanted to find out how well a neural machine translation could work with a language that didn't have millions of parallel lines (in fact, we have ~10,000). We also wanted to compare it to our rule-based transducer, which we've been making additions to since the end of the semester.

===Motivation===
This is the only (at least easily accessible when searching) machine translation available between Kaingang and another language. The motivation behind our project was to possibly allow the Kaingang people to not have to choose between teaching Kaingang or Portuguese to later generations. We hoped (if at some point one of the translators was state-of-the-art) any need for Portuguese could be accessed through the translator, which would be an incentive to keep the language Kaingang alive.

==What we did==
* Added many more features to and improved performance of rule-based transducer
* Trained a neural network with 75% of our parallel corpus and tested with 25% using OpenNMT
* Documented ambiguity in Kaingang
* Created parallel corpora for the whole new testament between Kaingang and Portuguese!
==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]

[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation Metric==
Our new evaluation accounts for the fact that words in the Apertium translation sentences has # and * symbols. The accuracy script looks something like this:
correct_word_percentage = []
for i in range(len(true_translations)):
correct_word_count = 0
for word in prediction[i].split():
if word.strip("\n").strip("*").strip("#") in true_translations[i]:
correct_word_count += 1
correct_word_percentage.append(correct_word_percentage/len(prediction[i].split()))
average_accuracy = sum(correct_word_percentage)/len(correct_word_percentage)
print(average_accuracy)
===Accuracy===
Initial RBMT accuracy:

Final RBMT accuracy: 9.91%

Initial NMT accuracy: 21.4%

Final NMT accuracy: (currently training on new larger corpus, so no final accuracy yet)

==Specs for rule-based==
===Number of...===
===Coverage===
* Coverage over new testament:
$ aq-covtest ../ling073-kgp-corpus/kgp.corpus.large.txt kgp.automorf.bin
Number of tokenised words in the corpus: 406362
Coverage: 92.93%
Top unknown words in the corpus:
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
270 nĩnh
269 tũꞌ
Translation time: 1.422363042831421 seconds
* Coverage over dictionary sentences:
$ aq-covtest ../ling073-kgp-por-corpus/kgp.sentences.txt kgp.automorf.bin
Number of tokenised words in the corpus: 32659
Coverage: 92.72%
Top unknown words in the corpus:
30 nĩm
27 vég
26 isóg
24 jagy
23 prũ
23 tãg
22 kyr
20 mỹnh
20 régre
18 pir
18 féj
17 jã
17 Kusa
17 nĩgé
16 kãtĩ
16 jamã
15 venh
15 gãnh
15 jakré
14 nug
Translation time: 0.1878523826599121 seconds

===Word Error Rate===

==Documented ambiguity in Kaingang==
ag - (pr pes) eles / (pr dem) os
ã -
ãjag -
e - (mod) muito / (v iv) fazer, causar
eg -
ẽmĩ - (n) pao de milho / (v iv) tatear, apalpar
fa - (n) perna, planta, amargo / (v tv sg) lavar roupa
fag - (pr pes) elas / (pr dem) as
fi - (pr pes) ela / (pr dem) a
ge - (mod) entao, tambem / (n) semelhante / (v iv) fazer igual
gé - (mod) tambem / (v tv pl) levar, carregar
gunhgunh - (n) picao preto / (v tv pl) fincar
han - (v iv sg) sarar, melhorar / (v tv sg) fazer
he - (ij) sim / (v tv) dizer
hur - (o) ja / (v iv sg) caimbra
hỹn - (o) de certo, provavelmente / (int) onde?
inh -
jãn -
jãvãnh - (mod) recusar, nao saber fazer / (v tv sg) esperar, procurar
jé - (su) futuro / (n) reza, hino
jẽn -
jo - (conj) entao / (cir) antes, na frente
jũn - (n) palmito / (v tv) embrabecer
kafãn -
ke - (n) sobra, resto / (v iv) futuro, fazer, dizer
ker - (o) cuidado / (n) presente de comida, cama de barbante de urtiga
kre - (n) balaio, coxa / (v tv pl) cortar, ceifar
kren - (mod) quase / (v iv sg) escapar / (v tv sg) perder
krenkren - (v iv pl) escapar / (v tv pl) perder
kur - (o) ligeiro, rapido! / (n) roupa, vestido
mé - (mod) fazer diariamente, ligeiro, gostar de fazer / (n) carneiro
mẽ - (mod) muito, ligeiro / (v tv sg) cheirar, escutar, sentir, tocar
mro - (v iv sg) tomar banho / (v tv sg) por de molho
mũ - (a) fazendo, narrativo, acao unica, consequencia / (v iv pl) andar
mỹ - (cir) para / (su) pergunta
mỹr - (o) verdadeiro, e certo / (conj) ?
ne - (su) originador / (v tv sg) enterrar / (n) enterrado / de / (int) o que?
nẽ - (a) sera que? / (v tv) cobrir encostado
nig - (n) lagoa, poco / (v tv sg) dar pontape
nĩ - (a) na situacao de, sentado, ter, ter a obrigacao, condicao / (n) carne / (v iv sg) sentar
nỹ - (a) deitado / (su) topico na pergunta / (n) mae, irma da mae / (v iv sg) deitar-se
nỹtĩ - (a) sendo, ter / (v iv pl) existir
pẽ - (mod) muito, o verdadeiro / (n) braco
ra - (cir) para, na direcao de, apesar de / (n) queixo / (a) faca ja!
rã - (cir) perto, por baixo de / (n) sol, maduro / (v iv sg) entrar, comecar, estar perto de
rán - (n) encosta, declive, barranco, subida, perau / (v tv sg) escrever
rãké - (v iv) chegando tarde / tarde ?
rãnhrãj - (n) trabalho / (v tv pl) trabalhar
re - (n) grama, campo, fora de casa / (v iv) encher-se de bebida / (v iv pl) descer
sa - (a) pendurado / (n) sal, cachoeira / (v iv) pendurado / (v tv) pendurar, pregar
se - (n) quati / (v tv sg) amarrar, prender / (ij) puxa!
tãnh - (n) morador, dono / (v iv) fazer forte, habituar-se
ti - (pr pes) ele / (pr dem) o
tógfĩn - (n) gaviao / (v tv sg) atar, amarrar
tỹ - (cir) com, por / (erg) / (ex) / (top)
vãn - (n) taquara / (v tv sg) carregar coisa comprida
vãnh - (mod) nao querer ou poder fazer algo / (n) capoeira, mato, plantas curtas
ve - (n) aparencia, natureza, irma de um homem, primeiro / (v iv sg) parecer ser / (v tv sg) ver, enxergar
vo - (v iv) barulho de andar no mato, coracao batendo / (n) macuco
vỹ - (su) topico / (o) sera que e? / (voc) homem! moco!

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T03:32:25Z

Drosset1: /* Documented ambiguity in Kaingang */

==The project: RBMT or NMT?==
===Background===
Kaingang is a language spoken in the southern Brazil region. It has roughly 20,000 speakers, many of whom are bilingual with Portuguese. Kaingang is one the most well-documented languages of indigenous peoples of Brazil; however, in our project, we were only able to use a [https://www.sil.org/resources/archives/42876 Kaingang-Portuguese dictionary] and a [https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf Kaingang new testament bible] (the other resources were written in languages we didn't read).
===Idea===
We wanted to find out how well a neural machine translation could work with a language that didn't have millions of parallel lines (in fact, we have ~10,000). We also wanted to compare it to our rule-based transducer, which we've been making additions to since the end of the semester.

===Motivation===
This is the only (at least easily accessible when searching) machine translation available between Kaingang and another language. The motivation behind our project was to possibly allow the Kaingang people to not have to choose between teaching Kaingang or Portuguese to later generations. We hoped (if at some point one of the translators was state-of-the-art) any need for Portuguese could be accessed through the translator, which would be an incentive to keep the language Kaingang alive.

==What we did==
* Added many more features to and improved performance of rule-based transducer
* Trained a neural network with 75% of our parallel corpus and tested with 25% using OpenNMT
* Documented ambiguity in Kaingang
* Created parallel corpora for the whole new testament between Kaingang and Portuguese!
==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]

[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation Metric==
Our new evaluation accounts for the fact that words in the Apertium translation sentences has # and * symbols. The accuracy script looks something like this:
correct_word_percentage = []
for i in range(len(true_translations)):
correct_word_count = 0
for word in prediction[i].split():
if word.strip("\n").strip("*").strip("#") in true_translations[i]:
correct_word_count += 1
correct_word_percentage.append(correct_word_percentage/len(prediction[i].split()))
average_accuracy = sum(correct_word_percentage)/len(correct_word_percentage)
print(average_accuracy)
===Accuracy===
Initial RBMT accuracy:

Final RBMT accuracy: 9.91%

Initial NMT accuracy: 21.4%

Final NMT accuracy: (currently training on new larger corpus, so no final accuracy yet)

==Specs for rule-based==
===Number of...===
===Coverage===
* Coverage over new testament:
$ aq-covtest ../ling073-kgp-corpus/kgp.corpus.large.txt kgp.automorf.bin
Number of tokenised words in the corpus: 406362
Coverage: 92.93%
Top unknown words in the corpus:
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
270 nĩnh
269 tũꞌ
Translation time: 1.422363042831421 seconds
* Coverage over dictionary sentences:
$ aq-covtest ../ling073-kgp-por-corpus/kgp.sentences.txt kgp.automorf.bin
Number of tokenised words in the corpus: 32659
Coverage: 92.72%
Top unknown words in the corpus:
30 nĩm
27 vég
26 isóg
24 jagy
23 prũ
23 tãg
22 kyr
20 mỹnh
20 régre
18 pir
18 féj
17 jã
17 Kusa
17 nĩgé
16 kãtĩ
16 jamã
15 venh
15 gãnh
15 jakré
14 nug
Translation time: 0.1878523826599121 seconds

===Word Error Rate===

==Documented ambiguity in Kaingang==
ag - (pr pes) eles / (pr dem) os
ã -
ãjag -
e - (mod) muito / (v iv) fazer, causar
eg -
ẽmĩ - (n) pao de milho / (v iv) tatear, apalpar
fa - (n) perna, planta, amargo / (v tv sg) lavar roupa
fag - (pr pes) elas / (pr dem) as
fi - (pr pes) ela / (pr dem) a
ge - (mod) entao, tambem / (n) semelhante / (v iv) fazer igual
gé - (mod) tambem / (v tv pl) levar, carregar
gunhgunh - (n) picao preto / (v tv pl) fincar
han - (v iv sg) sarar, melhorar / (v tv sg) fazer
he - (ij) sim / (v tv) dizer
hur - (o) ja / (v iv sg) caimbra
hỹn - (o) de certo, provavelmente / (int) onde?
inh -
jãn -
jãvãnh - (mod) recusar, nao saber fazer / (v tv sg) esperar, procurar
jé - (su) futuro / (n) reza, hino
jẽn -
jo - (conj) entao / (cir) antes, na frente
jũn - (n) palmito / (v tv) embrabecer
kafãn -
ke - (n) sobra, resto / (v iv) futuro, fazer, dizer
ker - (o) cuidado / (n) presente de comida, cama de barbante de urtiga
kre - (n) balaio, coxa / (v tv pl) cortar, ceifar
kren - (mod) quase / (v iv sg) escapar / (v tv sg) perder
krenkren - (v iv pl) escapar / (v tv pl) perder
kur - (o) ligeiro, rapido! / (n) roupa, vestido
mé - (mod) fazer diariamente, ligeiro, gostar de fazer / (n) carneiro
mẽ - (mod) muito, ligeiro / (v tv sg) cheirar, escutar, sentir, tocar
mro - (v iv sg) tomar banho / (v tv sg) por de molho
mũ - (a) fazendo, narrativo, acao unica, consequencia / (v iv pl) andar
mỹ - (cir) para / (su) pergunta
mỹr - (o) verdadeiro, e certo / (conj) ?
ne - (su) originador / (v tv sg) enterrar / (n) enterrado / de / (int) o que?
nẽ - (a) sera que? / (v tv) cobrir encostado
nig - (n) lagoa, poco / (v tv sg) dar pontape
nĩ - (a) na situacao de, sentado, ter, ter a obrigacao, condicao / (n) carne / (v iv sg) sentar
nỹ - (a) deitado / (su) topico na pergunta / (n) mae, irma da mae / (v iv sg) deitar-se
nỹtĩ - (a) sendo, ter / (v iv pl) existir
ón - not ambiguous actually
pẽ - (mod) muito, o verdadeiro / (n) braco
ra - (cir) para, na direcao de, apesar de / (n) queixo / (a) faca ja!
rã - (cir) perto, por baixo de / (n) sol, maduro / (v iv sg) entrar, comecar, estar perto de
rán - (n) encosta, declive, barranco, subida, perau / (v tv sg) escrever
rãké - (v iv) chegando tarde / tarde ?
rãnhrãj - (n) trabalho / (v tv pl) trabalhar
re - (n) grama, campo, fora de casa / (v iv) encher-se de bebida / (v iv pl) descer
sa - (a) pendurado / (n) sal, cachoeira / (v iv) pendurado / (v tv) pendurar, pregar
se - (n) quati / (v tv sg) amarrar, prender / (ij) puxa!
tãnh - (n) morador, dono / (v iv) fazer forte, habituar-se
ti - (pr pes) ele / (pr dem) o
to - not ambiguous?
tógfĩn - (n) gaviao / (v tv sg) atar, amarrar
tũ - not ambiguous?
tỹ - (cir) com, por / (erg) / (ex) / (top)
ũn - not ambiguous
vãn - (n) taquara / (v tv sg) carregar coisa comprida
vãnh - (mod) nao querer ou poder fazer algo / (n) capoeira, mato, plantas curtas
ve - (n) aparencia, natureza, irma de um homem, primeiro / (v iv sg) parecer ser / (v tv sg) ver, enxergar
vo - (v iv) barulho de andar no mato, coracao batendo / (n) macuco
vỹ - (su) topico / (o) sera que e? / (voc) homem! moco!

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T03:27:55Z

Drosset1: /* Accuracy */

Kaingang and Portuguese/Final Project

2019-05-15T03:27:25Z

Drosset1: /* Accuracy */

Kaingang and Portuguese/Final Project

2019-05-15T03:25:45Z

Drosset1: /* Coverage */

Kaingang and Portuguese/Final Project

2019-05-15T03:24:10Z

Drosset1: /* Coverage */

Kaingang and Portuguese/Final Project

2019-05-15T03:21:13Z

Drosset1:

Kaingang and Portuguese/Final Project

2019-05-15T03:16:49Z

Drosset1: /* Evaluation Metric */

Kaingang and Portuguese/Final Project

2019-05-15T03:16:41Z

Drosset1: /* Evaluation Metric */

Kaingang and Portuguese/Final Project

2019-05-15T03:16:32Z

Drosset1: /* Evaluation Metric */

Kaingang and Portuguese/Final Project

2019-05-15T03:16:19Z

Drosset1: /* Specs for rule-based */

Kaingang and Portuguese/Final Project

2019-05-15T03:15:46Z

Drosset1: /* Accuracy */

Kaingang and Portuguese/Final Project

2019-05-15T03:14:26Z

Drosset1: /* Evaluation Metric */

Kaingang and Portuguese/Final Project

2019-05-15T03:13:56Z

Drosset1: /* Evaluation */

Kaingang and Portuguese/Final Project

2019-05-15T03:08:38Z

Drosset1: /* Evaluation */

Kaingang and Portuguese/Final Project

2019-05-15T02:45:28Z

Drosset1: /* Idea */

Kaingang and Portuguese/Final Project

2019-05-15T02:45:12Z

Drosset1: /* The project: RBMT or NMT? */

Kaingang and Portuguese/Final Project

2019-05-15T02:30:20Z

Drosset1: /* The project */

==The project: RBMT or NMT?==
Kaingang is a language spoken in the southern Brazil region. It has roughly 20,000 speakers, many of whom are bilingual with Portuguese. Kaingang is one the most well-documented languages of indigenous peoples of Brazil; however, in our project, we were only able to use a [https://www.sil.org/resources/archives/42876 Kaingang-Portuguese dictionary] and a [https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf Kaingang new testament bible] (the other resources were written in languages we didn't read).

==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]

[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation==

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T02:17:08Z

Drosset1:

==The project==
==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]

[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation==

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T02:17:02Z

Drosset1:

==The project==
==Links to repositories==
[https://github.com/daniellerossetti/Kaingang-Transducer Kaingang transducer]
[https://github.com/daniellerossetti/Kaingang-Portuguese-Translator Kaingang-Portuguese translator]
==Evaluation==

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T02:15:59Z

Drosset1:

==The project==
==Links to repositories==

==Evaluation==

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese/Final Project

2019-05-15T02:13:44Z

Drosset1: Created page with "Category:Sp19 FinalProjects"

[[Category:Sp19 FinalProjects]]

Kaingang and Portuguese

2019-05-01T00:33:18Z

Drosset1: /* Additions for "Polished RBMT System" assignment */

Resources for machine translation between [[Kaingang]] and [[Portuguese]]. [https://github.com/erick-guti-prado/apertium-por]

==External Resources==

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por Repository for language pair]

*[https://github.com/erick-guti-prado/apertium-por Portuguese transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp Kaingang transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por-corpus Parallel corpus]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Contrastive_Grammar Contrastive Grammar between Kaingang and Portuguese]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Lexical_Selection Lexical selection]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Structural_Transfer Structural Transfer]

==kgp → por Evaluation==
===Coverage (number of forms analysed) of Kaingang transducer on kgp.sentences.txt file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 32669
Coverage: 89.98%
Top unknown words in the corpus:
160 há
60 kaga
52 mẽg
47 sĩ
46 si
44 nĩn
43 kanẽ
42 Ka
40 ka
39 ké
37 fe
34 pẽn
33 rá
30 nĩm
27 vég
27 panh
26 jafã
26 isóg
24 jagy
24 Jesus
Translation time: 0.08736515045166016 seconds

===Coverage of Kaingang-Portuguese transducer (xyz-abc.automorf.hfst) on the same file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp-por/kgp-por.automorf.bin
Number of tokenised words in the corpus: 34281
Coverage: 58.82%
Top unknown words in the corpus:
739 vỹ
300 ki
287 fi
280 to
198 há
189 sóg
188 vẽnh
182 tũ
171 tá
145 nĩgtĩ
135 tĩg
134 ha
127 mẽ
121 pẽ
114 kar
114 mág
111 nén
111 kã
105 jẽ
102 kri
Translation time: 0.06977176666259766 seconds
===10 sentences===
*'''Original sentence: ''' Ag mỹ tóg, e tĩ.
'''Intended Portuguese translation: ''' Ficaram surpresos.
'''kgp-por-biltrans output: ''' ^Ag<prn><pes>/Eles<prn><tn><pes>/Deles<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Eles para #ele, #fazer #tendo ido.
*'''Original sentence: ''' Ti mré ke ag vẽ.
'''Intended Portuguese translation: ''' São os seus companheiros.
'''kgp-por-biltrans output: ''' ^Ti<prn><pes>/Ele<prn><tn><pes>/Lhe<prn><tn><pes>/Seu<prn><tn><pes>/Dele<prn><tn><pes>$ ^mré ke<mod>/parente<n><m>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Ele #parente #eles #é.
*'''Original sentence: ''' Kã’ég ũ tóg sá nĩ, ũ tóg kusũg nĩ, ũ tóg tánh nĩ, ũ tóg kógnár nĩ, e ti nỹtĩ.
'''Intended Portuguese translation: ''' Algumas araras são pretas, outras vermelhas, outras verdeazuis, outras pintadas, há muitas variedades.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^sá<mod>/preto<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kusũg<mod>/vermelho<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tánh<mod>/azul<adj>/verde<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kógnár<mod>/pintado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^nỹtĩ<v><iv><pl>/existir<vblex>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Arara #alguém #ele #preto #sentar, #alguém #ele #vermelho #sentar, #alguém #ele #azul #sentar, #alguém #ele #pintado #sentar, #fazer #ele #existir
*'''Original sentence: ''' Tãnh jũn kym ẽg tĩ, kỹ ẽg tóg, ti fár tar ẽn vãvãm tĩ, jó ẽg ti fár tãnãj ẽn mré ti kãtá ke ẽn ko tĩ, tãnh kaju ti.
'''Intended Portuguese translation: ''' Cortamos o palmito, jogamos fora a casca dura e comemos a casca mole com o coração.
'''kgp-por-biltrans output: ''' ^jũn<n>/palmito<n><m>$ ^kym<v><tv><sg>/cortar<vblex>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^*tar/*tar$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^vãvãm<v><tv><pl>/vender<vblex>/descartar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^jó<cir>/antes<adv>/em frente<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^tãnãj<n>/mole<adj><mf>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^mré<cir>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^*kãtá/*kãtá$ ^ke<n>/sobra<n><f>/resto<n><m>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^ko<v><tv>/usar<vblex>/comer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^tãnh<n>/palmeira<n><f>$ ^kaju<v>/empurrar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Palmeira #palmito #cortar #nós #tendo ido, então #nós #ele, #ele #pele *tar #aquele #vender #tendo ido, #antes #nós #ele #pele #mole #aquele com #ele *kãtá #sobra #aquele #usar #tendo ido, #palmeira #empurrar #ele.
*'''Original sentence: ''' Kejẽn ũ tóg ẽg pãnĩ tá ẽg jé prẽr tĩ, kỹ ẽg tóg, vĩrĩn ke kỹ ti mĩ ẽvãnh kỹ, ti mré vĩ mũ sir.
'''Intended Portuguese translation: ''' Às vezes alguém nos chama nas nossas costas, aí nós nos viramos para olhá-lo e falar com ele.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^jé prẽr<v><sv>/chamar<vblex><sv>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^vĩrĩn ke<v><iv>/virar-se<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mĩ<cir>/dentro<adv>$ ^ẽvãnh<v><iv>/enxergar<vblex>/observar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mré<cir>/com<pr>$ ^vĩ<n>/palavra<n><f>/discurso<n><m>$ ^mũ<a>/fazendo<vblex>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #alguém #ele #nós atrás #nós #chamar #tendo ido, então #nós #ele, #virar-se então #ele dentro #enxergar então, #ele com #palavra #fazendo daí.
*'''Original sentence: '''Kejẽn ẽg tóg, re ra kutẽ tĩ, ẽg tỹ ĩn kãkã nĩ ra, kỹ ẽg tóg, jãnkã nón kỹ króm ke tĩ sir.
'''Intended Portuguese translation: ''' Às vezes quando saímos para fora da casa abrimos a porta e saímos.
'''kgp-por-biltrans output: ''' ^Kejẽn<cir>/Às vezes<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^re<n>/fora<adv>/grama<n><f>/campo<n><m>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^kutẽ<v><iv><sg>/cair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ĩn<n>/casa<n><f>$ ^kãkã<v><tv>/cavoucar<vblex>/revistar<vblex>/vasculhar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jãnkã<n>/porta<n><f>$ ^nón<cir>/após<pr>/atrás<adv>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^króm ke<v><iv>/sair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: '''Às vezes #nós #ele, fora para #cair #tendo ido, #nós por #casa #cavoucar #sentar para, então #nós #ele, #porta após então #sair #tendo ido daí.
*'''Original sentence: ''' Kejẽn ẽg japry tóg jonjor sag tĩ, kỹ ẽg tóg ti rãgfãn tĩ, ẽg tỹ vẽnhrã hã ti tĩn jé.
'''Intended Portuguese translation: ''' Às vezes o nosso caminho tem muitas curvas, então nós o endireitamos para ficar mais reto.
'''kgp-por-biltrans output: ''' ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^japry<n>/caminho# de algo<n><m>/caminho# de alguém<n><m>$ ^tóg<su>/ele<prn><tn>$ ^jonjor<n>/curvas<n><f>/ziguezague<n><m>$ ^sag<v><tv>/atar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^rãgfãn<v><tv>/endireitar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^vẽnh rã hã<cir>/perto<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^tĩn<v><tv>/fazer andar<vblex>/passar por cima de<vblex>$ ^jé<n>/reza<n><f>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #nós #caminho# de algo #ele #curvas #atar #tendo ido, então #nós #ele #ele #endireitar #tendo ido, #nós por perto #ele #fazer andar #reza.
*'''Original sentence: ''' “Inh kã’u ã,” he tóg. “Inh mỹ ã tóg tỹ, ũ nĩ ven nĩ vẽ, hã ra ã tóg tỹ, ã hã nĩ”, he tóg.
'''Intended Portuguese translation: ''' “Você me assustou, pensei que você era um outro mas é você mesmo”, ele falou para mim.
'''kgp-por-biltrans output: ''' “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^kã’u<v><tv>/assustar<vblex>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$^,<cm>/,<cm>$” ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$ “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ven<v><tv><sg>/mostrar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^,<cm>/,<cm>$ ^hã<o>/igual<adj>/parecido<adj>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^hã<o>/igual<adj>/parecido<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$”^,<cm>/,<cm>$ ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' “#Eu #assustar #tu,” sim #ele. “#Eu para #tu #ele por, #alguém #sentar #mostrar #sentar #é, #igual para #tu #ele por, #tu #igual #sentar”, sim #ele.
*'''Original sentence: ''' Kyfé tỹ ag tóg, jagnẽ mĩ tỹ kririn ke ja nĩ.
'''Intended Portuguese translation: ''' Eles haviam se riscado um ao outro com uma faca.
'''kgp-por-biltrans output: ''' ^tỹ<cir>/por<pr>/com<pr>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jagnẽ<prn>/si mesmo<prn>$ ^mĩ<cir>/dentro<adv>$ ^tỹ<cir>/por<pr>/com<pr>$ ^kririn ke<v><iv>/riscar<vblex>$ ^ja<mod>/feito<adj>/terminado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Faca por #eles #ele, #si mesmo dentro por #riscar #feito #sentar.
*'''Original sentence: ''' Ẽg tỹ ũ mũmẽg han kỹ ẽg tóg, ti pãnĩ tá ti vóg tĩ; ẽg tỹ ti vatánh vẽ.
'''Intended Portuguese translation: ''' Quando queremos dar um susto em alguém, nós o pegamos por trás, assim nós o assustamos.
'''kgp-por-biltrans output: ''' ^Ẽg<prn><pes>/Nós<prn><tn><pes>/Nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ũ<prn>/alguém<prn><tn>$ ^mũmẽg<v><iv>/estar com medo<vblex>$ ^han<v><iv><sg>/sarar<vblex>/melhorar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vóg<v><tv>/mexer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^;<sent>/;<sent>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vatánh<v><tv>/enganar<vblex>/assustar<vblex>/contrariar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Nós por #alguém #estar com medo #sarar então #nós #ele, #ele atrás #ele #mexer #tendo ido; #nós por #ele #enganar #é.

==Additions for "Polished RBMT System" assignment==
*Added 103 words to the bilingual dictionary
*Added 6 disambiguation rules
*Added 2 lexical transfer rules

==Final evaluation==
===kgp transducer===
*Precision and recall against annotated.basic corpus:
$ precisionRecall ../ling073-kgp-corpus/kgp.annotated.basic.txt ../ling073-kgp-corpus/kgp.annotated.basic.txt
Totals: 139 forms, 427 tp, 0 fp, 0 tn, 0 fn
Precision: 100.00000%
Recall: 100.00000%

*Coverage over the large corpus:
$ aq-covtest ling073-kgp-corpus/kgp.corpus.large.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 406588
Coverage: 91.99%
Top unknown words in the corpus:
726 nĩn
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
377 jafã
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
Translation time: 1.0561802387237549 seconds
*The number of words in the large corpus: 366,209
*The number of stems in the transducer: 2,299

===kgp-por pair===
*WER over longer corpus:
*PER over longer corpus:
*Proportion of stems translated correctly in the longer corpus:
*Trimmed coverage over longer and large corpora:
*The number of tokens in longer corpora:
*The number of tokens in large corpora: 406,588

[[Category:Sp19_TranslationPairs]] [[Category:Kaingang]] [[Category:Portuguese]]

Kaingang/Universal Dependencies

2019-05-01T00:16:30Z

Drosset1:

''kgp.annotated2.ud.conllu'' and ''kgp.annotated2.nomorph.conllu'' are the files we chose to train the parser on (not ''kgp.annotated.ud.conllu'' or ''kgp.annotated.nomorph.conllu'').

==Evaluation==
===Number of forms, UAS, LAS scores for ''kgp.annotated2.ud.conllu''===
$ udpipe --accuracy --parse kgp.withmorph.udpipe ../kgp.annotated2.ud.conllu
Loading UDPipe model: done.
Parsing from gold tokenization with gold tags - forms: 160, UAS: 70.62%, LAS: 55.00%
*Number of sentences in corpus: 18

===Number of forms, UAS, LAS scores for ''kgp.annotated2.nomorph.conllu''===
$ udpipe --accuracy --parse kgp.nomorph.udpipe ../kgp.annotated2.nomorph.conllu
Loading UDPipe model: done.
Parsing from gold tokenization with gold tags - forms: 160, UAS: 70.62%, LAS: 55.00%
*Number of sentences in corpus: 18

==Dependency==
(for each relation: a description of the relation, noting various ways it might be used in the language; and two examples of the relation from your corpus, preferably illustrating what you described.)
===Five dependency relations that we used at least twice in our annotation===
====case====
====nsubj====
====advmod====
====nmod:poss====
====obj====

[[Category: sp19_UD]] [[Category: Kaingang]]

Kaingang/Universal Dependencies

2019-04-30T23:59:24Z

Drosset1: /* Dependency */

Kaingang/Universal Dependencies

2019-04-30T23:55:17Z

Drosset1:

Kaingang/Universal Dependencies

2019-04-30T23:54:05Z

Drosset1: /* Evaluation */

==Evaluation==
===Number of forms, UAS, LAS scores for ''kgp.annotated2.ud.conllu''===
$ udpipe --accuracy --parse kgp.withmorph.udpipe ../kgp.annotated2.ud.conllu
Loading UDPipe model: done.
Parsing from gold tokenization with gold tags - forms: 160, UAS: 70.62%, LAS: 55.00%
*Number of sentences in corpus: 18

===Number of forms, UAS, LAS scores for ''kgp.annotated2.nomorph.conllu''===
$ udpipe --accuracy --parse kgp.nomorph.udpipe ../kgp.annotated2.nomorph.conllu
Loading UDPipe model: done.
Parsing from gold tokenization with gold tags - forms: 160, UAS: 70.62%, LAS: 55.00%
*Number of sentences in corpus: 18

==Dependency==

[[Category: sp19_UD]]

Kaingang and Portuguese

2019-04-30T23:49:21Z

Drosset1:

Resources for machine translation between [[Kaingang]] and [[Portuguese]]. [https://github.com/erick-guti-prado/apertium-por]

==External Resources==

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por Repository for language pair]

*[https://github.com/erick-guti-prado/apertium-por Portuguese transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp Kaingang transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por-corpus Parallel corpus]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Contrastive_Grammar Contrastive Grammar between Kaingang and Portuguese]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Lexical_Selection Lexical selection]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Structural_Transfer Structural Transfer]

==kgp → por Evaluation==
===Coverage (number of forms analysed) of Kaingang transducer on kgp.sentences.txt file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 32669
Coverage: 89.98%
Top unknown words in the corpus:
160 há
60 kaga
52 mẽg
47 sĩ
46 si
44 nĩn
43 kanẽ
42 Ka
40 ka
39 ké
37 fe
34 pẽn
33 rá
30 nĩm
27 vég
27 panh
26 jafã
26 isóg
24 jagy
24 Jesus
Translation time: 0.08736515045166016 seconds

===Coverage of Kaingang-Portuguese transducer (xyz-abc.automorf.hfst) on the same file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp-por/kgp-por.automorf.bin
Number of tokenised words in the corpus: 34281
Coverage: 58.82%
Top unknown words in the corpus:
739 vỹ
300 ki
287 fi
280 to
198 há
189 sóg
188 vẽnh
182 tũ
171 tá
145 nĩgtĩ
135 tĩg
134 ha
127 mẽ
121 pẽ
114 kar
114 mág
111 nén
111 kã
105 jẽ
102 kri
Translation time: 0.06977176666259766 seconds
===10 sentences===
*'''Original sentence: ''' Ag mỹ tóg, e tĩ.
'''Intended Portuguese translation: ''' Ficaram surpresos.
'''kgp-por-biltrans output: ''' ^Ag<prn><pes>/Eles<prn><tn><pes>/Deles<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Eles para #ele, #fazer #tendo ido.
*'''Original sentence: ''' Ti mré ke ag vẽ.
'''Intended Portuguese translation: ''' São os seus companheiros.
'''kgp-por-biltrans output: ''' ^Ti<prn><pes>/Ele<prn><tn><pes>/Lhe<prn><tn><pes>/Seu<prn><tn><pes>/Dele<prn><tn><pes>$ ^mré ke<mod>/parente<n><m>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Ele #parente #eles #é.
*'''Original sentence: ''' Kã’ég ũ tóg sá nĩ, ũ tóg kusũg nĩ, ũ tóg tánh nĩ, ũ tóg kógnár nĩ, e ti nỹtĩ.
'''Intended Portuguese translation: ''' Algumas araras são pretas, outras vermelhas, outras verdeazuis, outras pintadas, há muitas variedades.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^sá<mod>/preto<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kusũg<mod>/vermelho<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tánh<mod>/azul<adj>/verde<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kógnár<mod>/pintado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^nỹtĩ<v><iv><pl>/existir<vblex>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Arara #alguém #ele #preto #sentar, #alguém #ele #vermelho #sentar, #alguém #ele #azul #sentar, #alguém #ele #pintado #sentar, #fazer #ele #existir
*'''Original sentence: ''' Tãnh jũn kym ẽg tĩ, kỹ ẽg tóg, ti fár tar ẽn vãvãm tĩ, jó ẽg ti fár tãnãj ẽn mré ti kãtá ke ẽn ko tĩ, tãnh kaju ti.
'''Intended Portuguese translation: ''' Cortamos o palmito, jogamos fora a casca dura e comemos a casca mole com o coração.
'''kgp-por-biltrans output: ''' ^jũn<n>/palmito<n><m>$ ^kym<v><tv><sg>/cortar<vblex>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^*tar/*tar$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^vãvãm<v><tv><pl>/vender<vblex>/descartar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^jó<cir>/antes<adv>/em frente<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^tãnãj<n>/mole<adj><mf>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^mré<cir>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^*kãtá/*kãtá$ ^ke<n>/sobra<n><f>/resto<n><m>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^ko<v><tv>/usar<vblex>/comer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^tãnh<n>/palmeira<n><f>$ ^kaju<v>/empurrar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Palmeira #palmito #cortar #nós #tendo ido, então #nós #ele, #ele #pele *tar #aquele #vender #tendo ido, #antes #nós #ele #pele #mole #aquele com #ele *kãtá #sobra #aquele #usar #tendo ido, #palmeira #empurrar #ele.
*'''Original sentence: ''' Kejẽn ũ tóg ẽg pãnĩ tá ẽg jé prẽr tĩ, kỹ ẽg tóg, vĩrĩn ke kỹ ti mĩ ẽvãnh kỹ, ti mré vĩ mũ sir.
'''Intended Portuguese translation: ''' Às vezes alguém nos chama nas nossas costas, aí nós nos viramos para olhá-lo e falar com ele.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^jé prẽr<v><sv>/chamar<vblex><sv>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^vĩrĩn ke<v><iv>/virar-se<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mĩ<cir>/dentro<adv>$ ^ẽvãnh<v><iv>/enxergar<vblex>/observar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mré<cir>/com<pr>$ ^vĩ<n>/palavra<n><f>/discurso<n><m>$ ^mũ<a>/fazendo<vblex>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #alguém #ele #nós atrás #nós #chamar #tendo ido, então #nós #ele, #virar-se então #ele dentro #enxergar então, #ele com #palavra #fazendo daí.
*'''Original sentence: '''Kejẽn ẽg tóg, re ra kutẽ tĩ, ẽg tỹ ĩn kãkã nĩ ra, kỹ ẽg tóg, jãnkã nón kỹ króm ke tĩ sir.
'''Intended Portuguese translation: ''' Às vezes quando saímos para fora da casa abrimos a porta e saímos.
'''kgp-por-biltrans output: ''' ^Kejẽn<cir>/Às vezes<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^re<n>/fora<adv>/grama<n><f>/campo<n><m>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^kutẽ<v><iv><sg>/cair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ĩn<n>/casa<n><f>$ ^kãkã<v><tv>/cavoucar<vblex>/revistar<vblex>/vasculhar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jãnkã<n>/porta<n><f>$ ^nón<cir>/após<pr>/atrás<adv>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^króm ke<v><iv>/sair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: '''Às vezes #nós #ele, fora para #cair #tendo ido, #nós por #casa #cavoucar #sentar para, então #nós #ele, #porta após então #sair #tendo ido daí.
*'''Original sentence: ''' Kejẽn ẽg japry tóg jonjor sag tĩ, kỹ ẽg tóg ti rãgfãn tĩ, ẽg tỹ vẽnhrã hã ti tĩn jé.
'''Intended Portuguese translation: ''' Às vezes o nosso caminho tem muitas curvas, então nós o endireitamos para ficar mais reto.
'''kgp-por-biltrans output: ''' ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^japry<n>/caminho# de algo<n><m>/caminho# de alguém<n><m>$ ^tóg<su>/ele<prn><tn>$ ^jonjor<n>/curvas<n><f>/ziguezague<n><m>$ ^sag<v><tv>/atar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^rãgfãn<v><tv>/endireitar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^vẽnh rã hã<cir>/perto<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^tĩn<v><tv>/fazer andar<vblex>/passar por cima de<vblex>$ ^jé<n>/reza<n><f>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #nós #caminho# de algo #ele #curvas #atar #tendo ido, então #nós #ele #ele #endireitar #tendo ido, #nós por perto #ele #fazer andar #reza.
*'''Original sentence: ''' “Inh kã’u ã,” he tóg. “Inh mỹ ã tóg tỹ, ũ nĩ ven nĩ vẽ, hã ra ã tóg tỹ, ã hã nĩ”, he tóg.
'''Intended Portuguese translation: ''' “Você me assustou, pensei que você era um outro mas é você mesmo”, ele falou para mim.
'''kgp-por-biltrans output: ''' “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^kã’u<v><tv>/assustar<vblex>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$^,<cm>/,<cm>$” ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$ “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ven<v><tv><sg>/mostrar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^,<cm>/,<cm>$ ^hã<o>/igual<adj>/parecido<adj>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^hã<o>/igual<adj>/parecido<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$”^,<cm>/,<cm>$ ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' “#Eu #assustar #tu,” sim #ele. “#Eu para #tu #ele por, #alguém #sentar #mostrar #sentar #é, #igual para #tu #ele por, #tu #igual #sentar”, sim #ele.
*'''Original sentence: ''' Kyfé tỹ ag tóg, jagnẽ mĩ tỹ kririn ke ja nĩ.
'''Intended Portuguese translation: ''' Eles haviam se riscado um ao outro com uma faca.
'''kgp-por-biltrans output: ''' ^tỹ<cir>/por<pr>/com<pr>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jagnẽ<prn>/si mesmo<prn>$ ^mĩ<cir>/dentro<adv>$ ^tỹ<cir>/por<pr>/com<pr>$ ^kririn ke<v><iv>/riscar<vblex>$ ^ja<mod>/feito<adj>/terminado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Faca por #eles #ele, #si mesmo dentro por #riscar #feito #sentar.
*'''Original sentence: ''' Ẽg tỹ ũ mũmẽg han kỹ ẽg tóg, ti pãnĩ tá ti vóg tĩ; ẽg tỹ ti vatánh vẽ.
'''Intended Portuguese translation: ''' Quando queremos dar um susto em alguém, nós o pegamos por trás, assim nós o assustamos.
'''kgp-por-biltrans output: ''' ^Ẽg<prn><pes>/Nós<prn><tn><pes>/Nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ũ<prn>/alguém<prn><tn>$ ^mũmẽg<v><iv>/estar com medo<vblex>$ ^han<v><iv><sg>/sarar<vblex>/melhorar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vóg<v><tv>/mexer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^;<sent>/;<sent>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vatánh<v><tv>/enganar<vblex>/assustar<vblex>/contrariar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Nós por #alguém #estar com medo #sarar então #nós #ele, #ele atrás #ele #mexer #tendo ido; #nós por #ele #enganar #é.

==Additions for "Polished RBMT System" assignment==
*Added 103 words to the bilingual dictionary
*Added 6 disambiguation rules

==Final evaluation==
===kgp transducer===
*Precision and recall against annotated.basic corpus:
$ precisionRecall ../ling073-kgp-corpus/kgp.annotated.basic.txt ../ling073-kgp-corpus/kgp.annotated.basic.txt
Totals: 139 forms, 427 tp, 0 fp, 0 tn, 0 fn
Precision: 100.00000%
Recall: 100.00000%

*Coverage over the large corpus:
$ aq-covtest ling073-kgp-corpus/kgp.corpus.large.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 406588
Coverage: 91.99%
Top unknown words in the corpus:
726 nĩn
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
377 jafã
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
Translation time: 1.0561802387237549 seconds
*The number of words in the large corpus: 366,209
*The number of stems in the transducer: 2,299

===kgp-por pair===
*WER over longer corpus:
*PER over longer corpus:
*Proportion of stems translated correctly in the longer corpus:
*Trimmed coverage over longer and large corpora:
*The number of tokens in longer corpora:
*The number of tokens in large corpora: 406,588

[[Category:Sp19_TranslationPairs]] [[Category:Kaingang]] [[Category:Portuguese]]

Kaingang and Portuguese

2019-04-30T23:48:46Z

Drosset1:

Resources for machine translation between [[Kaingang]] and [[Portuguese]]. [https://github.com/erick-guti-prado/apertium-por]

==External Resources==

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por Repository for language pair]

*[https://github.com/erick-guti-prado/apertium-por Portuguese transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp Kaingang transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por-corpus Parallel corpus]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Contrastive_Grammar Contrastive Grammar between Kaingang and Portuguese]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Lexical_Selection Lexical selection]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Structural_Transfer Structural Transfer]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Universal_Dependencies Universal Dependencies]

==kgp → por Evaluation==
===Coverage (number of forms analysed) of Kaingang transducer on kgp.sentences.txt file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 32669
Coverage: 89.98%
Top unknown words in the corpus:
160 há
60 kaga
52 mẽg
47 sĩ
46 si
44 nĩn
43 kanẽ
42 Ka
40 ka
39 ké
37 fe
34 pẽn
33 rá
30 nĩm
27 vég
27 panh
26 jafã
26 isóg
24 jagy
24 Jesus
Translation time: 0.08736515045166016 seconds

===Coverage of Kaingang-Portuguese transducer (xyz-abc.automorf.hfst) on the same file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp-por/kgp-por.automorf.bin
Number of tokenised words in the corpus: 34281
Coverage: 58.82%
Top unknown words in the corpus:
739 vỹ
300 ki
287 fi
280 to
198 há
189 sóg
188 vẽnh
182 tũ
171 tá
145 nĩgtĩ
135 tĩg
134 ha
127 mẽ
121 pẽ
114 kar
114 mág
111 nén
111 kã
105 jẽ
102 kri
Translation time: 0.06977176666259766 seconds
===10 sentences===
*'''Original sentence: ''' Ag mỹ tóg, e tĩ.
'''Intended Portuguese translation: ''' Ficaram surpresos.
'''kgp-por-biltrans output: ''' ^Ag<prn><pes>/Eles<prn><tn><pes>/Deles<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Eles para #ele, #fazer #tendo ido.
*'''Original sentence: ''' Ti mré ke ag vẽ.
'''Intended Portuguese translation: ''' São os seus companheiros.
'''kgp-por-biltrans output: ''' ^Ti<prn><pes>/Ele<prn><tn><pes>/Lhe<prn><tn><pes>/Seu<prn><tn><pes>/Dele<prn><tn><pes>$ ^mré ke<mod>/parente<n><m>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Ele #parente #eles #é.
*'''Original sentence: ''' Kã’ég ũ tóg sá nĩ, ũ tóg kusũg nĩ, ũ tóg tánh nĩ, ũ tóg kógnár nĩ, e ti nỹtĩ.
'''Intended Portuguese translation: ''' Algumas araras são pretas, outras vermelhas, outras verdeazuis, outras pintadas, há muitas variedades.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^sá<mod>/preto<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kusũg<mod>/vermelho<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tánh<mod>/azul<adj>/verde<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kógnár<mod>/pintado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^nỹtĩ<v><iv><pl>/existir<vblex>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Arara #alguém #ele #preto #sentar, #alguém #ele #vermelho #sentar, #alguém #ele #azul #sentar, #alguém #ele #pintado #sentar, #fazer #ele #existir
*'''Original sentence: ''' Tãnh jũn kym ẽg tĩ, kỹ ẽg tóg, ti fár tar ẽn vãvãm tĩ, jó ẽg ti fár tãnãj ẽn mré ti kãtá ke ẽn ko tĩ, tãnh kaju ti.
'''Intended Portuguese translation: ''' Cortamos o palmito, jogamos fora a casca dura e comemos a casca mole com o coração.
'''kgp-por-biltrans output: ''' ^jũn<n>/palmito<n><m>$ ^kym<v><tv><sg>/cortar<vblex>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^*tar/*tar$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^vãvãm<v><tv><pl>/vender<vblex>/descartar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^jó<cir>/antes<adv>/em frente<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^tãnãj<n>/mole<adj><mf>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^mré<cir>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^*kãtá/*kãtá$ ^ke<n>/sobra<n><f>/resto<n><m>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^ko<v><tv>/usar<vblex>/comer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^tãnh<n>/palmeira<n><f>$ ^kaju<v>/empurrar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Palmeira #palmito #cortar #nós #tendo ido, então #nós #ele, #ele #pele *tar #aquele #vender #tendo ido, #antes #nós #ele #pele #mole #aquele com #ele *kãtá #sobra #aquele #usar #tendo ido, #palmeira #empurrar #ele.
*'''Original sentence: ''' Kejẽn ũ tóg ẽg pãnĩ tá ẽg jé prẽr tĩ, kỹ ẽg tóg, vĩrĩn ke kỹ ti mĩ ẽvãnh kỹ, ti mré vĩ mũ sir.
'''Intended Portuguese translation: ''' Às vezes alguém nos chama nas nossas costas, aí nós nos viramos para olhá-lo e falar com ele.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^jé prẽr<v><sv>/chamar<vblex><sv>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^vĩrĩn ke<v><iv>/virar-se<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mĩ<cir>/dentro<adv>$ ^ẽvãnh<v><iv>/enxergar<vblex>/observar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mré<cir>/com<pr>$ ^vĩ<n>/palavra<n><f>/discurso<n><m>$ ^mũ<a>/fazendo<vblex>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #alguém #ele #nós atrás #nós #chamar #tendo ido, então #nós #ele, #virar-se então #ele dentro #enxergar então, #ele com #palavra #fazendo daí.
*'''Original sentence: '''Kejẽn ẽg tóg, re ra kutẽ tĩ, ẽg tỹ ĩn kãkã nĩ ra, kỹ ẽg tóg, jãnkã nón kỹ króm ke tĩ sir.
'''Intended Portuguese translation: ''' Às vezes quando saímos para fora da casa abrimos a porta e saímos.
'''kgp-por-biltrans output: ''' ^Kejẽn<cir>/Às vezes<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^re<n>/fora<adv>/grama<n><f>/campo<n><m>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^kutẽ<v><iv><sg>/cair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ĩn<n>/casa<n><f>$ ^kãkã<v><tv>/cavoucar<vblex>/revistar<vblex>/vasculhar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jãnkã<n>/porta<n><f>$ ^nón<cir>/após<pr>/atrás<adv>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^króm ke<v><iv>/sair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: '''Às vezes #nós #ele, fora para #cair #tendo ido, #nós por #casa #cavoucar #sentar para, então #nós #ele, #porta após então #sair #tendo ido daí.
*'''Original sentence: ''' Kejẽn ẽg japry tóg jonjor sag tĩ, kỹ ẽg tóg ti rãgfãn tĩ, ẽg tỹ vẽnhrã hã ti tĩn jé.
'''Intended Portuguese translation: ''' Às vezes o nosso caminho tem muitas curvas, então nós o endireitamos para ficar mais reto.
'''kgp-por-biltrans output: ''' ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^japry<n>/caminho# de algo<n><m>/caminho# de alguém<n><m>$ ^tóg<su>/ele<prn><tn>$ ^jonjor<n>/curvas<n><f>/ziguezague<n><m>$ ^sag<v><tv>/atar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^rãgfãn<v><tv>/endireitar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^vẽnh rã hã<cir>/perto<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^tĩn<v><tv>/fazer andar<vblex>/passar por cima de<vblex>$ ^jé<n>/reza<n><f>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #nós #caminho# de algo #ele #curvas #atar #tendo ido, então #nós #ele #ele #endireitar #tendo ido, #nós por perto #ele #fazer andar #reza.
*'''Original sentence: ''' “Inh kã’u ã,” he tóg. “Inh mỹ ã tóg tỹ, ũ nĩ ven nĩ vẽ, hã ra ã tóg tỹ, ã hã nĩ”, he tóg.
'''Intended Portuguese translation: ''' “Você me assustou, pensei que você era um outro mas é você mesmo”, ele falou para mim.
'''kgp-por-biltrans output: ''' “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^kã’u<v><tv>/assustar<vblex>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$^,<cm>/,<cm>$” ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$ “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ven<v><tv><sg>/mostrar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^,<cm>/,<cm>$ ^hã<o>/igual<adj>/parecido<adj>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^hã<o>/igual<adj>/parecido<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$”^,<cm>/,<cm>$ ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' “#Eu #assustar #tu,” sim #ele. “#Eu para #tu #ele por, #alguém #sentar #mostrar #sentar #é, #igual para #tu #ele por, #tu #igual #sentar”, sim #ele.
*'''Original sentence: ''' Kyfé tỹ ag tóg, jagnẽ mĩ tỹ kririn ke ja nĩ.
'''Intended Portuguese translation: ''' Eles haviam se riscado um ao outro com uma faca.
'''kgp-por-biltrans output: ''' ^tỹ<cir>/por<pr>/com<pr>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jagnẽ<prn>/si mesmo<prn>$ ^mĩ<cir>/dentro<adv>$ ^tỹ<cir>/por<pr>/com<pr>$ ^kririn ke<v><iv>/riscar<vblex>$ ^ja<mod>/feito<adj>/terminado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Faca por #eles #ele, #si mesmo dentro por #riscar #feito #sentar.
*'''Original sentence: ''' Ẽg tỹ ũ mũmẽg han kỹ ẽg tóg, ti pãnĩ tá ti vóg tĩ; ẽg tỹ ti vatánh vẽ.
'''Intended Portuguese translation: ''' Quando queremos dar um susto em alguém, nós o pegamos por trás, assim nós o assustamos.
'''kgp-por-biltrans output: ''' ^Ẽg<prn><pes>/Nós<prn><tn><pes>/Nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ũ<prn>/alguém<prn><tn>$ ^mũmẽg<v><iv>/estar com medo<vblex>$ ^han<v><iv><sg>/sarar<vblex>/melhorar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vóg<v><tv>/mexer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^;<sent>/;<sent>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vatánh<v><tv>/enganar<vblex>/assustar<vblex>/contrariar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Nós por #alguém #estar com medo #sarar então #nós #ele, #ele atrás #ele #mexer #tendo ido; #nós por #ele #enganar #é.

==Additions for "Polished RBMT System" assignment==
*Added 103 words to the bilingual dictionary
*Added 6 disambiguation rules

==Final evaluation==
===kgp transducer===
*Precision and recall against annotated.basic corpus:
$ precisionRecall ../ling073-kgp-corpus/kgp.annotated.basic.txt ../ling073-kgp-corpus/kgp.annotated.basic.txt
Totals: 139 forms, 427 tp, 0 fp, 0 tn, 0 fn
Precision: 100.00000%
Recall: 100.00000%

*Coverage over the large corpus:
$ aq-covtest ling073-kgp-corpus/kgp.corpus.large.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 406588
Coverage: 91.99%
Top unknown words in the corpus:
726 nĩn
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
377 jafã
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
Translation time: 1.0561802387237549 seconds
*The number of words in the large corpus: 366,209
*The number of stems in the transducer: 2,299

===kgp-por pair===
*WER over longer corpus:
*PER over longer corpus:
*Proportion of stems translated correctly in the longer corpus:
*Trimmed coverage over longer and large corpora:
*The number of tokens in longer corpora:
*The number of tokens in large corpora: 406,588

[[Category:Sp19_TranslationPairs]] [[Category:Kaingang]] [[Category:Portuguese]]

Kaingang and Portuguese

2019-04-29T16:53:30Z

Drosset1: /* Additions */

Resources for machine translation between [[Kaingang]] and [[Portuguese]]. [https://github.com/erick-guti-prado/apertium-por]

==External Resources==

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por Repository for language pair]

*[https://github.com/erick-guti-prado/apertium-por Portuguese transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp Kaingang transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por-corpus Parallel corpus]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Contrastive_Grammar Contrastive Grammar between Kaingang and Portuguese]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Lexical_Selection Lexical selection]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Structural_Transfer Structural Transfer]

==kgp → por Evaluation==
===Coverage (number of forms analysed) of Kaingang transducer on kgp.sentences.txt file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 32669
Coverage: 89.98%
Top unknown words in the corpus:
160 há
60 kaga
52 mẽg
47 sĩ
46 si
44 nĩn
43 kanẽ
42 Ka
40 ka
39 ké
37 fe
34 pẽn
33 rá
30 nĩm
27 vég
27 panh
26 jafã
26 isóg
24 jagy
24 Jesus
Translation time: 0.08736515045166016 seconds

===Coverage of Kaingang-Portuguese transducer (xyz-abc.automorf.hfst) on the same file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp-por/kgp-por.automorf.bin
Number of tokenised words in the corpus: 34281
Coverage: 58.82%
Top unknown words in the corpus:
739 vỹ
300 ki
287 fi
280 to
198 há
189 sóg
188 vẽnh
182 tũ
171 tá
145 nĩgtĩ
135 tĩg
134 ha
127 mẽ
121 pẽ
114 kar
114 mág
111 nén
111 kã
105 jẽ
102 kri
Translation time: 0.06977176666259766 seconds
===10 sentences===
*'''Original sentence: ''' Ag mỹ tóg, e tĩ.
'''Intended Portuguese translation: ''' Ficaram surpresos.
'''kgp-por-biltrans output: ''' ^Ag<prn><pes>/Eles<prn><tn><pes>/Deles<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Eles para #ele, #fazer #tendo ido.
*'''Original sentence: ''' Ti mré ke ag vẽ.
'''Intended Portuguese translation: ''' São os seus companheiros.
'''kgp-por-biltrans output: ''' ^Ti<prn><pes>/Ele<prn><tn><pes>/Lhe<prn><tn><pes>/Seu<prn><tn><pes>/Dele<prn><tn><pes>$ ^mré ke<mod>/parente<n><m>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Ele #parente #eles #é.
*'''Original sentence: ''' Kã’ég ũ tóg sá nĩ, ũ tóg kusũg nĩ, ũ tóg tánh nĩ, ũ tóg kógnár nĩ, e ti nỹtĩ.
'''Intended Portuguese translation: ''' Algumas araras são pretas, outras vermelhas, outras verdeazuis, outras pintadas, há muitas variedades.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^sá<mod>/preto<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kusũg<mod>/vermelho<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tánh<mod>/azul<adj>/verde<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kógnár<mod>/pintado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^nỹtĩ<v><iv><pl>/existir<vblex>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Arara #alguém #ele #preto #sentar, #alguém #ele #vermelho #sentar, #alguém #ele #azul #sentar, #alguém #ele #pintado #sentar, #fazer #ele #existir
*'''Original sentence: ''' Tãnh jũn kym ẽg tĩ, kỹ ẽg tóg, ti fár tar ẽn vãvãm tĩ, jó ẽg ti fár tãnãj ẽn mré ti kãtá ke ẽn ko tĩ, tãnh kaju ti.
'''Intended Portuguese translation: ''' Cortamos o palmito, jogamos fora a casca dura e comemos a casca mole com o coração.
'''kgp-por-biltrans output: ''' ^jũn<n>/palmito<n><m>$ ^kym<v><tv><sg>/cortar<vblex>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^*tar/*tar$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^vãvãm<v><tv><pl>/vender<vblex>/descartar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^jó<cir>/antes<adv>/em frente<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^tãnãj<n>/mole<adj><mf>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^mré<cir>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^*kãtá/*kãtá$ ^ke<n>/sobra<n><f>/resto<n><m>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^ko<v><tv>/usar<vblex>/comer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^tãnh<n>/palmeira<n><f>$ ^kaju<v>/empurrar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Palmeira #palmito #cortar #nós #tendo ido, então #nós #ele, #ele #pele *tar #aquele #vender #tendo ido, #antes #nós #ele #pele #mole #aquele com #ele *kãtá #sobra #aquele #usar #tendo ido, #palmeira #empurrar #ele.
*'''Original sentence: ''' Kejẽn ũ tóg ẽg pãnĩ tá ẽg jé prẽr tĩ, kỹ ẽg tóg, vĩrĩn ke kỹ ti mĩ ẽvãnh kỹ, ti mré vĩ mũ sir.
'''Intended Portuguese translation: ''' Às vezes alguém nos chama nas nossas costas, aí nós nos viramos para olhá-lo e falar com ele.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^jé prẽr<v><sv>/chamar<vblex><sv>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^vĩrĩn ke<v><iv>/virar-se<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mĩ<cir>/dentro<adv>$ ^ẽvãnh<v><iv>/enxergar<vblex>/observar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mré<cir>/com<pr>$ ^vĩ<n>/palavra<n><f>/discurso<n><m>$ ^mũ<a>/fazendo<vblex>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #alguém #ele #nós atrás #nós #chamar #tendo ido, então #nós #ele, #virar-se então #ele dentro #enxergar então, #ele com #palavra #fazendo daí.
*'''Original sentence: '''Kejẽn ẽg tóg, re ra kutẽ tĩ, ẽg tỹ ĩn kãkã nĩ ra, kỹ ẽg tóg, jãnkã nón kỹ króm ke tĩ sir.
'''Intended Portuguese translation: ''' Às vezes quando saímos para fora da casa abrimos a porta e saímos.
'''kgp-por-biltrans output: ''' ^Kejẽn<cir>/Às vezes<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^re<n>/fora<adv>/grama<n><f>/campo<n><m>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^kutẽ<v><iv><sg>/cair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ĩn<n>/casa<n><f>$ ^kãkã<v><tv>/cavoucar<vblex>/revistar<vblex>/vasculhar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jãnkã<n>/porta<n><f>$ ^nón<cir>/após<pr>/atrás<adv>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^króm ke<v><iv>/sair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: '''Às vezes #nós #ele, fora para #cair #tendo ido, #nós por #casa #cavoucar #sentar para, então #nós #ele, #porta após então #sair #tendo ido daí.
*'''Original sentence: ''' Kejẽn ẽg japry tóg jonjor sag tĩ, kỹ ẽg tóg ti rãgfãn tĩ, ẽg tỹ vẽnhrã hã ti tĩn jé.
'''Intended Portuguese translation: ''' Às vezes o nosso caminho tem muitas curvas, então nós o endireitamos para ficar mais reto.
'''kgp-por-biltrans output: ''' ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^japry<n>/caminho# de algo<n><m>/caminho# de alguém<n><m>$ ^tóg<su>/ele<prn><tn>$ ^jonjor<n>/curvas<n><f>/ziguezague<n><m>$ ^sag<v><tv>/atar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^rãgfãn<v><tv>/endireitar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^vẽnh rã hã<cir>/perto<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^tĩn<v><tv>/fazer andar<vblex>/passar por cima de<vblex>$ ^jé<n>/reza<n><f>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #nós #caminho# de algo #ele #curvas #atar #tendo ido, então #nós #ele #ele #endireitar #tendo ido, #nós por perto #ele #fazer andar #reza.
*'''Original sentence: ''' “Inh kã’u ã,” he tóg. “Inh mỹ ã tóg tỹ, ũ nĩ ven nĩ vẽ, hã ra ã tóg tỹ, ã hã nĩ”, he tóg.
'''Intended Portuguese translation: ''' “Você me assustou, pensei que você era um outro mas é você mesmo”, ele falou para mim.
'''kgp-por-biltrans output: ''' “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^kã’u<v><tv>/assustar<vblex>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$^,<cm>/,<cm>$” ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$ “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ven<v><tv><sg>/mostrar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^,<cm>/,<cm>$ ^hã<o>/igual<adj>/parecido<adj>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^hã<o>/igual<adj>/parecido<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$”^,<cm>/,<cm>$ ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' “#Eu #assustar #tu,” sim #ele. “#Eu para #tu #ele por, #alguém #sentar #mostrar #sentar #é, #igual para #tu #ele por, #tu #igual #sentar”, sim #ele.
*'''Original sentence: ''' Kyfé tỹ ag tóg, jagnẽ mĩ tỹ kririn ke ja nĩ.
'''Intended Portuguese translation: ''' Eles haviam se riscado um ao outro com uma faca.
'''kgp-por-biltrans output: ''' ^tỹ<cir>/por<pr>/com<pr>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jagnẽ<prn>/si mesmo<prn>$ ^mĩ<cir>/dentro<adv>$ ^tỹ<cir>/por<pr>/com<pr>$ ^kririn ke<v><iv>/riscar<vblex>$ ^ja<mod>/feito<adj>/terminado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Faca por #eles #ele, #si mesmo dentro por #riscar #feito #sentar.
*'''Original sentence: ''' Ẽg tỹ ũ mũmẽg han kỹ ẽg tóg, ti pãnĩ tá ti vóg tĩ; ẽg tỹ ti vatánh vẽ.
'''Intended Portuguese translation: ''' Quando queremos dar um susto em alguém, nós o pegamos por trás, assim nós o assustamos.
'''kgp-por-biltrans output: ''' ^Ẽg<prn><pes>/Nós<prn><tn><pes>/Nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ũ<prn>/alguém<prn><tn>$ ^mũmẽg<v><iv>/estar com medo<vblex>$ ^han<v><iv><sg>/sarar<vblex>/melhorar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vóg<v><tv>/mexer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^;<sent>/;<sent>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vatánh<v><tv>/enganar<vblex>/assustar<vblex>/contrariar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Nós por #alguém #estar com medo #sarar então #nós #ele, #ele atrás #ele #mexer #tendo ido; #nós por #ele #enganar #é.

==Additions for "Polished RBMT System" assignment==
*Added 103 words to the bilingual dictionary
*Added 6 disambiguation rules

==Final evaluation==
===kgp transducer===
*Precision and recall against annotated.basic corpus:
$ precisionRecall ../ling073-kgp-corpus/kgp.annotated.basic.txt ../ling073-kgp-corpus/kgp.annotated.basic.txt
Totals: 139 forms, 427 tp, 0 fp, 0 tn, 0 fn
Precision: 100.00000%
Recall: 100.00000%

*Coverage over the large corpus:
$ aq-covtest ling073-kgp-corpus/kgp.corpus.large.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 406588
Coverage: 91.99%
Top unknown words in the corpus:
726 nĩn
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
377 jafã
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
Translation time: 1.0561802387237549 seconds
*The number of words in the large corpus: 366,209
*The number of stems in the transducer: 2,299

===kgp-por pair===
*WER over longer corpus:
*PER over longer corpus:
*Proportion of stems translated correctly in the longer corpus:
*Trimmed coverage over longer and large corpora:
*The number of tokens in longer corpora:
*The number of tokens in large corpora: 406,588

[[Category:Sp19_TranslationPairs]] [[Category:Kaingang]] [[Category:Portuguese]]

Kaingang

2019-04-25T13:34:14Z

Drosset1: /* Annotation */

The following is a documentation of Kaingang

==Developed resources==
===Keyboard===
[https://wikis.swarthmore.edu/ling073/Kaingang/Keyboard ''Kaingang keyboard layout'']

[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-keyboard ''Github repository for Kaingang keyboard layout'']

===Corpora===
[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-corpus ''A few corpora of text in Kaingang'']

===Grammar===
[https://wikis.swarthmore.edu/ling073/Kaingang/Grammar ''Kaingang grammar'']

===Morphological Transducer===
[https://wikis.swarthmore.edu/ling073/Kaingang/Transducer ''Kaingang Transducer wiki'']

[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp ''Link to Kaingang Transducer'']

[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese ''Resources for machine translation between Kaingang and Portuguese'']

[https://wikis.swarthmore.edu/ling073/Kaingang/Disambiguation ''Notes on disambiguation'']

===Translation===

[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese ''Resources for machine translation between Kaingang and Portuguese'']

[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por-corpus ''Kaingang and Portuguese parallel corpora'']

[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Contrastive_Grammar ''Differences in grammar between Kaingang and Portuguese'']

===Annotation===
[https://wikis.swarthmore.edu/ling073/Kaingang/Universal_Dependencies ''Universal Dependencies'']

==External Resources==
===Computational Resources===

[http://www.silbrasil.org.br/recursos/teclado_indigena_brasileiro '' Brazilian "Indigenous" Keyboard (a keyboard with characters commonly used in Brazilian indigenous languages)'']

===Dictionaries/Phrasebooks/Glossaries===

[https://www.sil.org/resources/archives/42876 ''Portuguese-Kaingang Dictionary'']

===Grammatical Descriptions===

[www.portalkaingang.org/lgua_kaingang.pdf ''Short overview of speakers, orthography, and semantics (PDF)'']

[http://www.academia.edu/241935/Fonologia_e_pros%C3%B3dia_do_Kaing%C3%A1ng_falado_em_Cacique_Doble_RS_ ''Phonological description'']

[http://www.omniglot.com/writing/kaingang.php ''Kaingang alphabet and another phonological description'']

[https://amerindias.github.io/curso2015/referencias/wie86pronounmacroje.pdf ''Pronoun system in Kaingang'']

[www.etnolinguistica.org/local--files/tese...2007/Goncalves_SolangeAparecida_M.pdf ''Description of aspect (PDF)'']

[https://sites.google.com/site/marcelojolkesky/relativas.pdf?attredirects=0 ''Relative phrases in Kaingang'']

[https://www.jstor.org/stable/25841469 ''Time distinctions in Kaingang'']

[https://wals.info/languoid/lect/wals_code_kng ''Some documented features of Kaingang (phonology, nominal categories, nominal syntax, verbal categories, word order, simple causes, and lexicon) and their sources'']

[https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&ved=2ahUKEwjGmorphLLgAhVOh-AKHWT6CrYQFjABegQICBAC&url=http%3A%2F%2Fwww.etnolinguistica.org%2Flocal--files%2Fbiblio%3Awiesemann-1972%2FWiesemann_1972_Struktur_der_Kaingang-Sprache.pdf&usg=AOvVaw1766caqecD2dVGfDeJ0pin ''Wieselmann 1972 Comprehensive Grammar'']

[http://ezinearticles.com/?Ergativity-is-Still-Alive-in-Kaingang-Language-From-South-Brazil-Part-2&id=491271 ''Ergativity is still alive in Kaingang'']

===Scientific Works===

[https://www.sil.org/resources/archives/6252 ''Inquiry into feasibility of standardization among regional dialects'']

[https://sites.google.com/site/marcelojolkesky/semi%C3%B3tica.pdf?attredirects=0 ''Semiotic analysis'']

===Corpora (any collection of authentic text, linguistically annotated or not, including Wikipedia in the language, news websites, Twitter feeds, etc.)===

[https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf ''Another bible in Kaingang'']

[https://www.bible.com/pt/bible/629/jhn.1.kgpnt ''New Testament'']

[http://www.marysrosaries.com/Kaingang_prayers.html '''Our Father' prayer'']

[https://www.youtube.com/watch?v=Q2AKqNkuFwI ''YouTube video of Kaingang woman singing a Brazilian song in Kaingang'']

[http://kanhgag.org/ ''Kaingang Webpage'']

[[Category:sp19_ResourceDocumentation]] [[Category:Kaingang]]

Kaingang

2019-04-25T13:34:02Z

Drosset1: /* Developed resources */

The following is a documentation of Kaingang

==Developed resources==
===Keyboard===
[https://wikis.swarthmore.edu/ling073/Kaingang/Keyboard ''Kaingang keyboard layout'']

[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-keyboard ''Github repository for Kaingang keyboard layout'']

===Corpora===
[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-corpus ''A few corpora of text in Kaingang'']

===Grammar===
[https://wikis.swarthmore.edu/ling073/Kaingang/Grammar ''Kaingang grammar'']

===Morphological Transducer===
[https://wikis.swarthmore.edu/ling073/Kaingang/Transducer ''Kaingang Transducer wiki'']

[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp ''Link to Kaingang Transducer'']

[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese ''Resources for machine translation between Kaingang and Portuguese'']

[https://wikis.swarthmore.edu/ling073/Kaingang/Disambiguation ''Notes on disambiguation'']

===Translation===

[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese ''Resources for machine translation between Kaingang and Portuguese'']

[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por-corpus ''Kaingang and Portuguese parallel corpora'']

[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Contrastive_Grammar ''Differences in grammar between Kaingang and Portuguese'']

==Annotation==
[https://wikis.swarthmore.edu/ling073/Kaingang/Universal_Dependencies ''Universal Dependencies'']

==External Resources==
===Computational Resources===

[http://www.silbrasil.org.br/recursos/teclado_indigena_brasileiro '' Brazilian "Indigenous" Keyboard (a keyboard with characters commonly used in Brazilian indigenous languages)'']

===Dictionaries/Phrasebooks/Glossaries===

[https://www.sil.org/resources/archives/42876 ''Portuguese-Kaingang Dictionary'']

===Grammatical Descriptions===

[www.portalkaingang.org/lgua_kaingang.pdf ''Short overview of speakers, orthography, and semantics (PDF)'']

[http://www.academia.edu/241935/Fonologia_e_pros%C3%B3dia_do_Kaing%C3%A1ng_falado_em_Cacique_Doble_RS_ ''Phonological description'']

[http://www.omniglot.com/writing/kaingang.php ''Kaingang alphabet and another phonological description'']

[https://amerindias.github.io/curso2015/referencias/wie86pronounmacroje.pdf ''Pronoun system in Kaingang'']

[www.etnolinguistica.org/local--files/tese...2007/Goncalves_SolangeAparecida_M.pdf ''Description of aspect (PDF)'']

[https://sites.google.com/site/marcelojolkesky/relativas.pdf?attredirects=0 ''Relative phrases in Kaingang'']

[https://www.jstor.org/stable/25841469 ''Time distinctions in Kaingang'']

[https://wals.info/languoid/lect/wals_code_kng ''Some documented features of Kaingang (phonology, nominal categories, nominal syntax, verbal categories, word order, simple causes, and lexicon) and their sources'']

[https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=2&ved=2ahUKEwjGmorphLLgAhVOh-AKHWT6CrYQFjABegQICBAC&url=http%3A%2F%2Fwww.etnolinguistica.org%2Flocal--files%2Fbiblio%3Awiesemann-1972%2FWiesemann_1972_Struktur_der_Kaingang-Sprache.pdf&usg=AOvVaw1766caqecD2dVGfDeJ0pin ''Wieselmann 1972 Comprehensive Grammar'']

[http://ezinearticles.com/?Ergativity-is-Still-Alive-in-Kaingang-Language-From-South-Brazil-Part-2&id=491271 ''Ergativity is still alive in Kaingang'']

===Scientific Works===

[https://www.sil.org/resources/archives/6252 ''Inquiry into feasibility of standardization among regional dialects'']

[https://sites.google.com/site/marcelojolkesky/semi%C3%B3tica.pdf?attredirects=0 ''Semiotic analysis'']

===Corpora (any collection of authentic text, linguistically annotated or not, including Wikipedia in the language, news websites, Twitter feeds, etc.)===

[https://s3.amazonaws.com/scriptureearth/data/kgp/PDF/00-WNTkgp-web.pdf ''Another bible in Kaingang'']

[https://www.bible.com/pt/bible/629/jhn.1.kgpnt ''New Testament'']

[http://www.marysrosaries.com/Kaingang_prayers.html '''Our Father' prayer'']

[https://www.youtube.com/watch?v=Q2AKqNkuFwI ''YouTube video of Kaingang woman singing a Brazilian song in Kaingang'']

[http://kanhgag.org/ ''Kaingang Webpage'']

[[Category:sp19_ResourceDocumentation]] [[Category:Kaingang]]

Kaingang and Portuguese

2019-04-23T13:14:46Z

Drosset1: /* kgp-por pair */

Resources for machine translation between [[Kaingang]] and [[Portuguese]]. [https://github.com/erick-guti-prado/apertium-por]

==External Resources==

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por Repository for language pair]

*[https://github.com/erick-guti-prado/apertium-por Portuguese transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp Kaingang transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por-corpus Parallel corpus]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Contrastive_Grammar Contrastive Grammar between Kaingang and Portuguese]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Lexical_Selection Lexical selection]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Structural_Transfer Structural Transfer]

==kgp → por Evaluation==
===Coverage (number of forms analysed) of Kaingang transducer on kgp.sentences.txt file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 32669
Coverage: 89.98%
Top unknown words in the corpus:
160 há
60 kaga
52 mẽg
47 sĩ
46 si
44 nĩn
43 kanẽ
42 Ka
40 ka
39 ké
37 fe
34 pẽn
33 rá
30 nĩm
27 vég
27 panh
26 jafã
26 isóg
24 jagy
24 Jesus
Translation time: 0.08736515045166016 seconds

===Coverage of Kaingang-Portuguese transducer (xyz-abc.automorf.hfst) on the same file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp-por/kgp-por.automorf.bin
Number of tokenised words in the corpus: 34281
Coverage: 58.82%
Top unknown words in the corpus:
739 vỹ
300 ki
287 fi
280 to
198 há
189 sóg
188 vẽnh
182 tũ
171 tá
145 nĩgtĩ
135 tĩg
134 ha
127 mẽ
121 pẽ
114 kar
114 mág
111 nén
111 kã
105 jẽ
102 kri
Translation time: 0.06977176666259766 seconds
===10 sentences===
*'''Original sentence: ''' Ag mỹ tóg, e tĩ.
'''Intended Portuguese translation: ''' Ficaram surpresos.
'''kgp-por-biltrans output: ''' ^Ag<prn><pes>/Eles<prn><tn><pes>/Deles<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Eles para #ele, #fazer #tendo ido.
*'''Original sentence: ''' Ti mré ke ag vẽ.
'''Intended Portuguese translation: ''' São os seus companheiros.
'''kgp-por-biltrans output: ''' ^Ti<prn><pes>/Ele<prn><tn><pes>/Lhe<prn><tn><pes>/Seu<prn><tn><pes>/Dele<prn><tn><pes>$ ^mré ke<mod>/parente<n><m>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Ele #parente #eles #é.
*'''Original sentence: ''' Kã’ég ũ tóg sá nĩ, ũ tóg kusũg nĩ, ũ tóg tánh nĩ, ũ tóg kógnár nĩ, e ti nỹtĩ.
'''Intended Portuguese translation: ''' Algumas araras são pretas, outras vermelhas, outras verdeazuis, outras pintadas, há muitas variedades.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^sá<mod>/preto<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kusũg<mod>/vermelho<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tánh<mod>/azul<adj>/verde<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kógnár<mod>/pintado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^nỹtĩ<v><iv><pl>/existir<vblex>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Arara #alguém #ele #preto #sentar, #alguém #ele #vermelho #sentar, #alguém #ele #azul #sentar, #alguém #ele #pintado #sentar, #fazer #ele #existir
*'''Original sentence: ''' Tãnh jũn kym ẽg tĩ, kỹ ẽg tóg, ti fár tar ẽn vãvãm tĩ, jó ẽg ti fár tãnãj ẽn mré ti kãtá ke ẽn ko tĩ, tãnh kaju ti.
'''Intended Portuguese translation: ''' Cortamos o palmito, jogamos fora a casca dura e comemos a casca mole com o coração.
'''kgp-por-biltrans output: ''' ^jũn<n>/palmito<n><m>$ ^kym<v><tv><sg>/cortar<vblex>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^*tar/*tar$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^vãvãm<v><tv><pl>/vender<vblex>/descartar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^jó<cir>/antes<adv>/em frente<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^tãnãj<n>/mole<adj><mf>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^mré<cir>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^*kãtá/*kãtá$ ^ke<n>/sobra<n><f>/resto<n><m>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^ko<v><tv>/usar<vblex>/comer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^tãnh<n>/palmeira<n><f>$ ^kaju<v>/empurrar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Palmeira #palmito #cortar #nós #tendo ido, então #nós #ele, #ele #pele *tar #aquele #vender #tendo ido, #antes #nós #ele #pele #mole #aquele com #ele *kãtá #sobra #aquele #usar #tendo ido, #palmeira #empurrar #ele.
*'''Original sentence: ''' Kejẽn ũ tóg ẽg pãnĩ tá ẽg jé prẽr tĩ, kỹ ẽg tóg, vĩrĩn ke kỹ ti mĩ ẽvãnh kỹ, ti mré vĩ mũ sir.
'''Intended Portuguese translation: ''' Às vezes alguém nos chama nas nossas costas, aí nós nos viramos para olhá-lo e falar com ele.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^jé prẽr<v><sv>/chamar<vblex><sv>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^vĩrĩn ke<v><iv>/virar-se<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mĩ<cir>/dentro<adv>$ ^ẽvãnh<v><iv>/enxergar<vblex>/observar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mré<cir>/com<pr>$ ^vĩ<n>/palavra<n><f>/discurso<n><m>$ ^mũ<a>/fazendo<vblex>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #alguém #ele #nós atrás #nós #chamar #tendo ido, então #nós #ele, #virar-se então #ele dentro #enxergar então, #ele com #palavra #fazendo daí.
*'''Original sentence: '''Kejẽn ẽg tóg, re ra kutẽ tĩ, ẽg tỹ ĩn kãkã nĩ ra, kỹ ẽg tóg, jãnkã nón kỹ króm ke tĩ sir.
'''Intended Portuguese translation: ''' Às vezes quando saímos para fora da casa abrimos a porta e saímos.
'''kgp-por-biltrans output: ''' ^Kejẽn<cir>/Às vezes<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^re<n>/fora<adv>/grama<n><f>/campo<n><m>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^kutẽ<v><iv><sg>/cair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ĩn<n>/casa<n><f>$ ^kãkã<v><tv>/cavoucar<vblex>/revistar<vblex>/vasculhar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jãnkã<n>/porta<n><f>$ ^nón<cir>/após<pr>/atrás<adv>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^króm ke<v><iv>/sair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: '''Às vezes #nós #ele, fora para #cair #tendo ido, #nós por #casa #cavoucar #sentar para, então #nós #ele, #porta após então #sair #tendo ido daí.
*'''Original sentence: ''' Kejẽn ẽg japry tóg jonjor sag tĩ, kỹ ẽg tóg ti rãgfãn tĩ, ẽg tỹ vẽnhrã hã ti tĩn jé.
'''Intended Portuguese translation: ''' Às vezes o nosso caminho tem muitas curvas, então nós o endireitamos para ficar mais reto.
'''kgp-por-biltrans output: ''' ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^japry<n>/caminho# de algo<n><m>/caminho# de alguém<n><m>$ ^tóg<su>/ele<prn><tn>$ ^jonjor<n>/curvas<n><f>/ziguezague<n><m>$ ^sag<v><tv>/atar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^rãgfãn<v><tv>/endireitar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^vẽnh rã hã<cir>/perto<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^tĩn<v><tv>/fazer andar<vblex>/passar por cima de<vblex>$ ^jé<n>/reza<n><f>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #nós #caminho# de algo #ele #curvas #atar #tendo ido, então #nós #ele #ele #endireitar #tendo ido, #nós por perto #ele #fazer andar #reza.
*'''Original sentence: ''' “Inh kã’u ã,” he tóg. “Inh mỹ ã tóg tỹ, ũ nĩ ven nĩ vẽ, hã ra ã tóg tỹ, ã hã nĩ”, he tóg.
'''Intended Portuguese translation: ''' “Você me assustou, pensei que você era um outro mas é você mesmo”, ele falou para mim.
'''kgp-por-biltrans output: ''' “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^kã’u<v><tv>/assustar<vblex>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$^,<cm>/,<cm>$” ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$ “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ven<v><tv><sg>/mostrar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^,<cm>/,<cm>$ ^hã<o>/igual<adj>/parecido<adj>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^hã<o>/igual<adj>/parecido<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$”^,<cm>/,<cm>$ ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' “#Eu #assustar #tu,” sim #ele. “#Eu para #tu #ele por, #alguém #sentar #mostrar #sentar #é, #igual para #tu #ele por, #tu #igual #sentar”, sim #ele.
*'''Original sentence: ''' Kyfé tỹ ag tóg, jagnẽ mĩ tỹ kririn ke ja nĩ.
'''Intended Portuguese translation: ''' Eles haviam se riscado um ao outro com uma faca.
'''kgp-por-biltrans output: ''' ^tỹ<cir>/por<pr>/com<pr>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jagnẽ<prn>/si mesmo<prn>$ ^mĩ<cir>/dentro<adv>$ ^tỹ<cir>/por<pr>/com<pr>$ ^kririn ke<v><iv>/riscar<vblex>$ ^ja<mod>/feito<adj>/terminado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Faca por #eles #ele, #si mesmo dentro por #riscar #feito #sentar.
*'''Original sentence: ''' Ẽg tỹ ũ mũmẽg han kỹ ẽg tóg, ti pãnĩ tá ti vóg tĩ; ẽg tỹ ti vatánh vẽ.
'''Intended Portuguese translation: ''' Quando queremos dar um susto em alguém, nós o pegamos por trás, assim nós o assustamos.
'''kgp-por-biltrans output: ''' ^Ẽg<prn><pes>/Nós<prn><tn><pes>/Nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ũ<prn>/alguém<prn><tn>$ ^mũmẽg<v><iv>/estar com medo<vblex>$ ^han<v><iv><sg>/sarar<vblex>/melhorar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vóg<v><tv>/mexer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^;<sent>/;<sent>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vatánh<v><tv>/enganar<vblex>/assustar<vblex>/contrariar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Nós por #alguém #estar com medo #sarar então #nós #ele, #ele atrás #ele #mexer #tendo ido; #nós por #ele #enganar #é.

==Additions==
*Added 103 words to the bilingual dictionary for "Polished RBMT System" assignment
==Final evaluation==
===kgp transducer===
*Precision and recall against annotated.basic corpus:
$ precisionRecall ../ling073-kgp-corpus/kgp.annotated.basic.txt ../ling073-kgp-corpus/kgp.annotated.basic.txt
Totals: 139 forms, 427 tp, 0 fp, 0 tn, 0 fn
Precision: 100.00000%
Recall: 100.00000%

*Coverage over the large corpus:
$ aq-covtest ling073-kgp-corpus/kgp.corpus.large.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 406588
Coverage: 91.99%
Top unknown words in the corpus:
726 nĩn
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
377 jafã
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
Translation time: 1.0561802387237549 seconds
*The number of words in the large corpus: 366,209
*The number of stems in the transducer: 2,299

===kgp-por pair===
*WER over longer corpus:
*PER over longer corpus:
*Proportion of stems translated correctly in the longer corpus:
*Trimmed coverage over longer and large corpora:
*The number of tokens in longer corpora:
*The number of tokens in large corpora: 406,588

[[Category:Sp19_TranslationPairs]] [[Category:Kaingang]] [[Category:Portuguese]]

Kaingang and Portuguese

2019-04-23T12:42:21Z

Drosset1: /* kgp transducer */

Resources for machine translation between [[Kaingang]] and [[Portuguese]]. [https://github.com/erick-guti-prado/apertium-por]

==External Resources==

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por Repository for language pair]

*[https://github.com/erick-guti-prado/apertium-por Portuguese transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp Kaingang transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por-corpus Parallel corpus]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Contrastive_Grammar Contrastive Grammar between Kaingang and Portuguese]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Lexical_Selection Lexical selection]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Structural_Transfer Structural Transfer]

==kgp → por Evaluation==
===Coverage (number of forms analysed) of Kaingang transducer on kgp.sentences.txt file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 32669
Coverage: 89.98%
Top unknown words in the corpus:
160 há
60 kaga
52 mẽg
47 sĩ
46 si
44 nĩn
43 kanẽ
42 Ka
40 ka
39 ké
37 fe
34 pẽn
33 rá
30 nĩm
27 vég
27 panh
26 jafã
26 isóg
24 jagy
24 Jesus
Translation time: 0.08736515045166016 seconds

===Coverage of Kaingang-Portuguese transducer (xyz-abc.automorf.hfst) on the same file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp-por/kgp-por.automorf.bin
Number of tokenised words in the corpus: 34281
Coverage: 58.82%
Top unknown words in the corpus:
739 vỹ
300 ki
287 fi
280 to
198 há
189 sóg
188 vẽnh
182 tũ
171 tá
145 nĩgtĩ
135 tĩg
134 ha
127 mẽ
121 pẽ
114 kar
114 mág
111 nén
111 kã
105 jẽ
102 kri
Translation time: 0.06977176666259766 seconds
===10 sentences===
*'''Original sentence: ''' Ag mỹ tóg, e tĩ.
'''Intended Portuguese translation: ''' Ficaram surpresos.
'''kgp-por-biltrans output: ''' ^Ag<prn><pes>/Eles<prn><tn><pes>/Deles<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Eles para #ele, #fazer #tendo ido.
*'''Original sentence: ''' Ti mré ke ag vẽ.
'''Intended Portuguese translation: ''' São os seus companheiros.
'''kgp-por-biltrans output: ''' ^Ti<prn><pes>/Ele<prn><tn><pes>/Lhe<prn><tn><pes>/Seu<prn><tn><pes>/Dele<prn><tn><pes>$ ^mré ke<mod>/parente<n><m>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Ele #parente #eles #é.
*'''Original sentence: ''' Kã’ég ũ tóg sá nĩ, ũ tóg kusũg nĩ, ũ tóg tánh nĩ, ũ tóg kógnár nĩ, e ti nỹtĩ.
'''Intended Portuguese translation: ''' Algumas araras são pretas, outras vermelhas, outras verdeazuis, outras pintadas, há muitas variedades.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^sá<mod>/preto<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kusũg<mod>/vermelho<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tánh<mod>/azul<adj>/verde<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kógnár<mod>/pintado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^nỹtĩ<v><iv><pl>/existir<vblex>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Arara #alguém #ele #preto #sentar, #alguém #ele #vermelho #sentar, #alguém #ele #azul #sentar, #alguém #ele #pintado #sentar, #fazer #ele #existir
*'''Original sentence: ''' Tãnh jũn kym ẽg tĩ, kỹ ẽg tóg, ti fár tar ẽn vãvãm tĩ, jó ẽg ti fár tãnãj ẽn mré ti kãtá ke ẽn ko tĩ, tãnh kaju ti.
'''Intended Portuguese translation: ''' Cortamos o palmito, jogamos fora a casca dura e comemos a casca mole com o coração.
'''kgp-por-biltrans output: ''' ^jũn<n>/palmito<n><m>$ ^kym<v><tv><sg>/cortar<vblex>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^*tar/*tar$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^vãvãm<v><tv><pl>/vender<vblex>/descartar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^jó<cir>/antes<adv>/em frente<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^tãnãj<n>/mole<adj><mf>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^mré<cir>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^*kãtá/*kãtá$ ^ke<n>/sobra<n><f>/resto<n><m>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^ko<v><tv>/usar<vblex>/comer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^tãnh<n>/palmeira<n><f>$ ^kaju<v>/empurrar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Palmeira #palmito #cortar #nós #tendo ido, então #nós #ele, #ele #pele *tar #aquele #vender #tendo ido, #antes #nós #ele #pele #mole #aquele com #ele *kãtá #sobra #aquele #usar #tendo ido, #palmeira #empurrar #ele.
*'''Original sentence: ''' Kejẽn ũ tóg ẽg pãnĩ tá ẽg jé prẽr tĩ, kỹ ẽg tóg, vĩrĩn ke kỹ ti mĩ ẽvãnh kỹ, ti mré vĩ mũ sir.
'''Intended Portuguese translation: ''' Às vezes alguém nos chama nas nossas costas, aí nós nos viramos para olhá-lo e falar com ele.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^jé prẽr<v><sv>/chamar<vblex><sv>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^vĩrĩn ke<v><iv>/virar-se<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mĩ<cir>/dentro<adv>$ ^ẽvãnh<v><iv>/enxergar<vblex>/observar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mré<cir>/com<pr>$ ^vĩ<n>/palavra<n><f>/discurso<n><m>$ ^mũ<a>/fazendo<vblex>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #alguém #ele #nós atrás #nós #chamar #tendo ido, então #nós #ele, #virar-se então #ele dentro #enxergar então, #ele com #palavra #fazendo daí.
*'''Original sentence: '''Kejẽn ẽg tóg, re ra kutẽ tĩ, ẽg tỹ ĩn kãkã nĩ ra, kỹ ẽg tóg, jãnkã nón kỹ króm ke tĩ sir.
'''Intended Portuguese translation: ''' Às vezes quando saímos para fora da casa abrimos a porta e saímos.
'''kgp-por-biltrans output: ''' ^Kejẽn<cir>/Às vezes<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^re<n>/fora<adv>/grama<n><f>/campo<n><m>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^kutẽ<v><iv><sg>/cair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ĩn<n>/casa<n><f>$ ^kãkã<v><tv>/cavoucar<vblex>/revistar<vblex>/vasculhar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jãnkã<n>/porta<n><f>$ ^nón<cir>/após<pr>/atrás<adv>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^króm ke<v><iv>/sair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: '''Às vezes #nós #ele, fora para #cair #tendo ido, #nós por #casa #cavoucar #sentar para, então #nós #ele, #porta após então #sair #tendo ido daí.
*'''Original sentence: ''' Kejẽn ẽg japry tóg jonjor sag tĩ, kỹ ẽg tóg ti rãgfãn tĩ, ẽg tỹ vẽnhrã hã ti tĩn jé.
'''Intended Portuguese translation: ''' Às vezes o nosso caminho tem muitas curvas, então nós o endireitamos para ficar mais reto.
'''kgp-por-biltrans output: ''' ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^japry<n>/caminho# de algo<n><m>/caminho# de alguém<n><m>$ ^tóg<su>/ele<prn><tn>$ ^jonjor<n>/curvas<n><f>/ziguezague<n><m>$ ^sag<v><tv>/atar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^rãgfãn<v><tv>/endireitar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^vẽnh rã hã<cir>/perto<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^tĩn<v><tv>/fazer andar<vblex>/passar por cima de<vblex>$ ^jé<n>/reza<n><f>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #nós #caminho# de algo #ele #curvas #atar #tendo ido, então #nós #ele #ele #endireitar #tendo ido, #nós por perto #ele #fazer andar #reza.
*'''Original sentence: ''' “Inh kã’u ã,” he tóg. “Inh mỹ ã tóg tỹ, ũ nĩ ven nĩ vẽ, hã ra ã tóg tỹ, ã hã nĩ”, he tóg.
'''Intended Portuguese translation: ''' “Você me assustou, pensei que você era um outro mas é você mesmo”, ele falou para mim.
'''kgp-por-biltrans output: ''' “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^kã’u<v><tv>/assustar<vblex>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$^,<cm>/,<cm>$” ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$ “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ven<v><tv><sg>/mostrar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^,<cm>/,<cm>$ ^hã<o>/igual<adj>/parecido<adj>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^hã<o>/igual<adj>/parecido<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$”^,<cm>/,<cm>$ ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' “#Eu #assustar #tu,” sim #ele. “#Eu para #tu #ele por, #alguém #sentar #mostrar #sentar #é, #igual para #tu #ele por, #tu #igual #sentar”, sim #ele.
*'''Original sentence: ''' Kyfé tỹ ag tóg, jagnẽ mĩ tỹ kririn ke ja nĩ.
'''Intended Portuguese translation: ''' Eles haviam se riscado um ao outro com uma faca.
'''kgp-por-biltrans output: ''' ^tỹ<cir>/por<pr>/com<pr>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jagnẽ<prn>/si mesmo<prn>$ ^mĩ<cir>/dentro<adv>$ ^tỹ<cir>/por<pr>/com<pr>$ ^kririn ke<v><iv>/riscar<vblex>$ ^ja<mod>/feito<adj>/terminado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Faca por #eles #ele, #si mesmo dentro por #riscar #feito #sentar.
*'''Original sentence: ''' Ẽg tỹ ũ mũmẽg han kỹ ẽg tóg, ti pãnĩ tá ti vóg tĩ; ẽg tỹ ti vatánh vẽ.
'''Intended Portuguese translation: ''' Quando queremos dar um susto em alguém, nós o pegamos por trás, assim nós o assustamos.
'''kgp-por-biltrans output: ''' ^Ẽg<prn><pes>/Nós<prn><tn><pes>/Nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ũ<prn>/alguém<prn><tn>$ ^mũmẽg<v><iv>/estar com medo<vblex>$ ^han<v><iv><sg>/sarar<vblex>/melhorar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vóg<v><tv>/mexer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^;<sent>/;<sent>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vatánh<v><tv>/enganar<vblex>/assustar<vblex>/contrariar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Nós por #alguém #estar com medo #sarar então #nós #ele, #ele atrás #ele #mexer #tendo ido; #nós por #ele #enganar #é.

==Additions==
*Added 103 words to the bilingual dictionary for "Polished RBMT System" assignment
==Final evaluation==
===kgp transducer===
*Precision and recall against annotated.basic corpus:
$ precisionRecall ../ling073-kgp-corpus/kgp.annotated.basic.txt ../ling073-kgp-corpus/kgp.annotated.basic.txt
Totals: 139 forms, 427 tp, 0 fp, 0 tn, 0 fn
Precision: 100.00000%
Recall: 100.00000%

*Coverage over the large corpus:
$ aq-covtest ling073-kgp-corpus/kgp.corpus.large.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 406588
Coverage: 91.99%
Top unknown words in the corpus:
726 nĩn
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
377 jafã
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
Translation time: 1.0561802387237549 seconds
*The number of words in the large corpus: 366,209
*The number of stems in the transducer: 2,299

===kgp-por pair===
*WER over longer corpus:
*PER over longer corpus:
*Proportion of stems translated correctly in the longer corpus:
*Trimmed coverage over longer and large corpora:
*The number of tokens in longer and large corpora:

[[Category:Sp19_TranslationPairs]] [[Category:Kaingang]] [[Category:Portuguese]]

Kaingang and Portuguese

2019-04-23T12:42:09Z

Drosset1: /* Final evaluation */

Resources for machine translation between [[Kaingang]] and [[Portuguese]]. [https://github.com/erick-guti-prado/apertium-por]

==External Resources==

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por Repository for language pair]

*[https://github.com/erick-guti-prado/apertium-por Portuguese transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp Kaingang transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por-corpus Parallel corpus]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Contrastive_Grammar Contrastive Grammar between Kaingang and Portuguese]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Lexical_Selection Lexical selection]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Structural_Transfer Structural Transfer]

==kgp → por Evaluation==
===Coverage (number of forms analysed) of Kaingang transducer on kgp.sentences.txt file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 32669
Coverage: 89.98%
Top unknown words in the corpus:
160 há
60 kaga
52 mẽg
47 sĩ
46 si
44 nĩn
43 kanẽ
42 Ka
40 ka
39 ké
37 fe
34 pẽn
33 rá
30 nĩm
27 vég
27 panh
26 jafã
26 isóg
24 jagy
24 Jesus
Translation time: 0.08736515045166016 seconds

===Coverage of Kaingang-Portuguese transducer (xyz-abc.automorf.hfst) on the same file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp-por/kgp-por.automorf.bin
Number of tokenised words in the corpus: 34281
Coverage: 58.82%
Top unknown words in the corpus:
739 vỹ
300 ki
287 fi
280 to
198 há
189 sóg
188 vẽnh
182 tũ
171 tá
145 nĩgtĩ
135 tĩg
134 ha
127 mẽ
121 pẽ
114 kar
114 mág
111 nén
111 kã
105 jẽ
102 kri
Translation time: 0.06977176666259766 seconds
===10 sentences===
*'''Original sentence: ''' Ag mỹ tóg, e tĩ.
'''Intended Portuguese translation: ''' Ficaram surpresos.
'''kgp-por-biltrans output: ''' ^Ag<prn><pes>/Eles<prn><tn><pes>/Deles<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Eles para #ele, #fazer #tendo ido.
*'''Original sentence: ''' Ti mré ke ag vẽ.
'''Intended Portuguese translation: ''' São os seus companheiros.
'''kgp-por-biltrans output: ''' ^Ti<prn><pes>/Ele<prn><tn><pes>/Lhe<prn><tn><pes>/Seu<prn><tn><pes>/Dele<prn><tn><pes>$ ^mré ke<mod>/parente<n><m>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Ele #parente #eles #é.
*'''Original sentence: ''' Kã’ég ũ tóg sá nĩ, ũ tóg kusũg nĩ, ũ tóg tánh nĩ, ũ tóg kógnár nĩ, e ti nỹtĩ.
'''Intended Portuguese translation: ''' Algumas araras são pretas, outras vermelhas, outras verdeazuis, outras pintadas, há muitas variedades.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^sá<mod>/preto<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kusũg<mod>/vermelho<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tánh<mod>/azul<adj>/verde<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kógnár<mod>/pintado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^nỹtĩ<v><iv><pl>/existir<vblex>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Arara #alguém #ele #preto #sentar, #alguém #ele #vermelho #sentar, #alguém #ele #azul #sentar, #alguém #ele #pintado #sentar, #fazer #ele #existir
*'''Original sentence: ''' Tãnh jũn kym ẽg tĩ, kỹ ẽg tóg, ti fár tar ẽn vãvãm tĩ, jó ẽg ti fár tãnãj ẽn mré ti kãtá ke ẽn ko tĩ, tãnh kaju ti.
'''Intended Portuguese translation: ''' Cortamos o palmito, jogamos fora a casca dura e comemos a casca mole com o coração.
'''kgp-por-biltrans output: ''' ^jũn<n>/palmito<n><m>$ ^kym<v><tv><sg>/cortar<vblex>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^*tar/*tar$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^vãvãm<v><tv><pl>/vender<vblex>/descartar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^jó<cir>/antes<adv>/em frente<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^tãnãj<n>/mole<adj><mf>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^mré<cir>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^*kãtá/*kãtá$ ^ke<n>/sobra<n><f>/resto<n><m>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^ko<v><tv>/usar<vblex>/comer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^tãnh<n>/palmeira<n><f>$ ^kaju<v>/empurrar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Palmeira #palmito #cortar #nós #tendo ido, então #nós #ele, #ele #pele *tar #aquele #vender #tendo ido, #antes #nós #ele #pele #mole #aquele com #ele *kãtá #sobra #aquele #usar #tendo ido, #palmeira #empurrar #ele.
*'''Original sentence: ''' Kejẽn ũ tóg ẽg pãnĩ tá ẽg jé prẽr tĩ, kỹ ẽg tóg, vĩrĩn ke kỹ ti mĩ ẽvãnh kỹ, ti mré vĩ mũ sir.
'''Intended Portuguese translation: ''' Às vezes alguém nos chama nas nossas costas, aí nós nos viramos para olhá-lo e falar com ele.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^jé prẽr<v><sv>/chamar<vblex><sv>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^vĩrĩn ke<v><iv>/virar-se<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mĩ<cir>/dentro<adv>$ ^ẽvãnh<v><iv>/enxergar<vblex>/observar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mré<cir>/com<pr>$ ^vĩ<n>/palavra<n><f>/discurso<n><m>$ ^mũ<a>/fazendo<vblex>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #alguém #ele #nós atrás #nós #chamar #tendo ido, então #nós #ele, #virar-se então #ele dentro #enxergar então, #ele com #palavra #fazendo daí.
*'''Original sentence: '''Kejẽn ẽg tóg, re ra kutẽ tĩ, ẽg tỹ ĩn kãkã nĩ ra, kỹ ẽg tóg, jãnkã nón kỹ króm ke tĩ sir.
'''Intended Portuguese translation: ''' Às vezes quando saímos para fora da casa abrimos a porta e saímos.
'''kgp-por-biltrans output: ''' ^Kejẽn<cir>/Às vezes<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^re<n>/fora<adv>/grama<n><f>/campo<n><m>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^kutẽ<v><iv><sg>/cair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ĩn<n>/casa<n><f>$ ^kãkã<v><tv>/cavoucar<vblex>/revistar<vblex>/vasculhar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jãnkã<n>/porta<n><f>$ ^nón<cir>/após<pr>/atrás<adv>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^króm ke<v><iv>/sair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: '''Às vezes #nós #ele, fora para #cair #tendo ido, #nós por #casa #cavoucar #sentar para, então #nós #ele, #porta após então #sair #tendo ido daí.
*'''Original sentence: ''' Kejẽn ẽg japry tóg jonjor sag tĩ, kỹ ẽg tóg ti rãgfãn tĩ, ẽg tỹ vẽnhrã hã ti tĩn jé.
'''Intended Portuguese translation: ''' Às vezes o nosso caminho tem muitas curvas, então nós o endireitamos para ficar mais reto.
'''kgp-por-biltrans output: ''' ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^japry<n>/caminho# de algo<n><m>/caminho# de alguém<n><m>$ ^tóg<su>/ele<prn><tn>$ ^jonjor<n>/curvas<n><f>/ziguezague<n><m>$ ^sag<v><tv>/atar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^rãgfãn<v><tv>/endireitar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^vẽnh rã hã<cir>/perto<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^tĩn<v><tv>/fazer andar<vblex>/passar por cima de<vblex>$ ^jé<n>/reza<n><f>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #nós #caminho# de algo #ele #curvas #atar #tendo ido, então #nós #ele #ele #endireitar #tendo ido, #nós por perto #ele #fazer andar #reza.
*'''Original sentence: ''' “Inh kã’u ã,” he tóg. “Inh mỹ ã tóg tỹ, ũ nĩ ven nĩ vẽ, hã ra ã tóg tỹ, ã hã nĩ”, he tóg.
'''Intended Portuguese translation: ''' “Você me assustou, pensei que você era um outro mas é você mesmo”, ele falou para mim.
'''kgp-por-biltrans output: ''' “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^kã’u<v><tv>/assustar<vblex>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$^,<cm>/,<cm>$” ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$ “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ven<v><tv><sg>/mostrar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^,<cm>/,<cm>$ ^hã<o>/igual<adj>/parecido<adj>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^hã<o>/igual<adj>/parecido<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$”^,<cm>/,<cm>$ ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' “#Eu #assustar #tu,” sim #ele. “#Eu para #tu #ele por, #alguém #sentar #mostrar #sentar #é, #igual para #tu #ele por, #tu #igual #sentar”, sim #ele.
*'''Original sentence: ''' Kyfé tỹ ag tóg, jagnẽ mĩ tỹ kririn ke ja nĩ.
'''Intended Portuguese translation: ''' Eles haviam se riscado um ao outro com uma faca.
'''kgp-por-biltrans output: ''' ^tỹ<cir>/por<pr>/com<pr>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jagnẽ<prn>/si mesmo<prn>$ ^mĩ<cir>/dentro<adv>$ ^tỹ<cir>/por<pr>/com<pr>$ ^kririn ke<v><iv>/riscar<vblex>$ ^ja<mod>/feito<adj>/terminado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Faca por #eles #ele, #si mesmo dentro por #riscar #feito #sentar.
*'''Original sentence: ''' Ẽg tỹ ũ mũmẽg han kỹ ẽg tóg, ti pãnĩ tá ti vóg tĩ; ẽg tỹ ti vatánh vẽ.
'''Intended Portuguese translation: ''' Quando queremos dar um susto em alguém, nós o pegamos por trás, assim nós o assustamos.
'''kgp-por-biltrans output: ''' ^Ẽg<prn><pes>/Nós<prn><tn><pes>/Nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ũ<prn>/alguém<prn><tn>$ ^mũmẽg<v><iv>/estar com medo<vblex>$ ^han<v><iv><sg>/sarar<vblex>/melhorar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vóg<v><tv>/mexer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^;<sent>/;<sent>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vatánh<v><tv>/enganar<vblex>/assustar<vblex>/contrariar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Nós por #alguém #estar com medo #sarar então #nós #ele, #ele atrás #ele #mexer #tendo ido; #nós por #ele #enganar #é.

==Additions==
*Added 103 words to the bilingual dictionary for "Polished RBMT System" assignment
==Final evaluation==
===kgp transducer===
*Precision and recall against annotated.basic corpus:
$ precisionRecall ../ling073-kgp-corpus/kgp.annotated.basic.txt ../ling073-kgp-corpus/kgp.annotated.basic.txt
Totals: 139 forms, 427 tp, 0 fp, 0 tn, 0 fn
Precision: 100.00000%
Recall: 100.00000%

*Coverage over the large corpus:
$ aq-covtest ling073-kgp-corpus/kgp.corpus.large.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 406588
Coverage: 91.99%
Top unknown words in the corpus:
726 nĩn
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
377 jafã
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
Translation time: 1.0561802387237549 seconds
*The number of words in the large corpus: 366,209
*The number of stems in the transducer: 2,299

===kgp-por pair===
*WER over longer corpus:
*PER over longer corpus:
*Proportion of stems translated correctly in the longer corpus:
*Trimmed coverage over longer and large corpora:
*The number of tokens in longer and large corpora:

[[Category:Sp19_TranslationPairs]] [[Category:Kaingang]] [[Category:Portuguese]]

Kaingang and Portuguese

2019-04-20T04:57:46Z

Drosset1: /* Final evaluation */

Resources for machine translation between [[Kaingang]] and [[Portuguese]]. [https://github.com/erick-guti-prado/apertium-por]

==External Resources==

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por Repository for language pair]

*[https://github.com/erick-guti-prado/apertium-por Portuguese transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp Kaingang transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por-corpus Parallel corpus]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Contrastive_Grammar Contrastive Grammar between Kaingang and Portuguese]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Lexical_Selection Lexical selection]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Structural_Transfer Structural Transfer]

==kgp → por Evaluation==
===Coverage (number of forms analysed) of Kaingang transducer on kgp.sentences.txt file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 32669
Coverage: 89.98%
Top unknown words in the corpus:
160 há
60 kaga
52 mẽg
47 sĩ
46 si
44 nĩn
43 kanẽ
42 Ka
40 ka
39 ké
37 fe
34 pẽn
33 rá
30 nĩm
27 vég
27 panh
26 jafã
26 isóg
24 jagy
24 Jesus
Translation time: 0.08736515045166016 seconds

===Coverage of Kaingang-Portuguese transducer (xyz-abc.automorf.hfst) on the same file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp-por/kgp-por.automorf.bin
Number of tokenised words in the corpus: 34281
Coverage: 58.82%
Top unknown words in the corpus:
739 vỹ
300 ki
287 fi
280 to
198 há
189 sóg
188 vẽnh
182 tũ
171 tá
145 nĩgtĩ
135 tĩg
134 ha
127 mẽ
121 pẽ
114 kar
114 mág
111 nén
111 kã
105 jẽ
102 kri
Translation time: 0.06977176666259766 seconds
===10 sentences===
*'''Original sentence: ''' Ag mỹ tóg, e tĩ.
'''Intended Portuguese translation: ''' Ficaram surpresos.
'''kgp-por-biltrans output: ''' ^Ag<prn><pes>/Eles<prn><tn><pes>/Deles<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Eles para #ele, #fazer #tendo ido.
*'''Original sentence: ''' Ti mré ke ag vẽ.
'''Intended Portuguese translation: ''' São os seus companheiros.
'''kgp-por-biltrans output: ''' ^Ti<prn><pes>/Ele<prn><tn><pes>/Lhe<prn><tn><pes>/Seu<prn><tn><pes>/Dele<prn><tn><pes>$ ^mré ke<mod>/parente<n><m>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Ele #parente #eles #é.
*'''Original sentence: ''' Kã’ég ũ tóg sá nĩ, ũ tóg kusũg nĩ, ũ tóg tánh nĩ, ũ tóg kógnár nĩ, e ti nỹtĩ.
'''Intended Portuguese translation: ''' Algumas araras são pretas, outras vermelhas, outras verdeazuis, outras pintadas, há muitas variedades.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^sá<mod>/preto<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kusũg<mod>/vermelho<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tánh<mod>/azul<adj>/verde<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kógnár<mod>/pintado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^nỹtĩ<v><iv><pl>/existir<vblex>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Arara #alguém #ele #preto #sentar, #alguém #ele #vermelho #sentar, #alguém #ele #azul #sentar, #alguém #ele #pintado #sentar, #fazer #ele #existir
*'''Original sentence: ''' Tãnh jũn kym ẽg tĩ, kỹ ẽg tóg, ti fár tar ẽn vãvãm tĩ, jó ẽg ti fár tãnãj ẽn mré ti kãtá ke ẽn ko tĩ, tãnh kaju ti.
'''Intended Portuguese translation: ''' Cortamos o palmito, jogamos fora a casca dura e comemos a casca mole com o coração.
'''kgp-por-biltrans output: ''' ^jũn<n>/palmito<n><m>$ ^kym<v><tv><sg>/cortar<vblex>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^*tar/*tar$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^vãvãm<v><tv><pl>/vender<vblex>/descartar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^jó<cir>/antes<adv>/em frente<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^tãnãj<n>/mole<adj><mf>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^mré<cir>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^*kãtá/*kãtá$ ^ke<n>/sobra<n><f>/resto<n><m>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^ko<v><tv>/usar<vblex>/comer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^tãnh<n>/palmeira<n><f>$ ^kaju<v>/empurrar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Palmeira #palmito #cortar #nós #tendo ido, então #nós #ele, #ele #pele *tar #aquele #vender #tendo ido, #antes #nós #ele #pele #mole #aquele com #ele *kãtá #sobra #aquele #usar #tendo ido, #palmeira #empurrar #ele.
*'''Original sentence: ''' Kejẽn ũ tóg ẽg pãnĩ tá ẽg jé prẽr tĩ, kỹ ẽg tóg, vĩrĩn ke kỹ ti mĩ ẽvãnh kỹ, ti mré vĩ mũ sir.
'''Intended Portuguese translation: ''' Às vezes alguém nos chama nas nossas costas, aí nós nos viramos para olhá-lo e falar com ele.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^jé prẽr<v><sv>/chamar<vblex><sv>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^vĩrĩn ke<v><iv>/virar-se<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mĩ<cir>/dentro<adv>$ ^ẽvãnh<v><iv>/enxergar<vblex>/observar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mré<cir>/com<pr>$ ^vĩ<n>/palavra<n><f>/discurso<n><m>$ ^mũ<a>/fazendo<vblex>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #alguém #ele #nós atrás #nós #chamar #tendo ido, então #nós #ele, #virar-se então #ele dentro #enxergar então, #ele com #palavra #fazendo daí.
*'''Original sentence: '''Kejẽn ẽg tóg, re ra kutẽ tĩ, ẽg tỹ ĩn kãkã nĩ ra, kỹ ẽg tóg, jãnkã nón kỹ króm ke tĩ sir.
'''Intended Portuguese translation: ''' Às vezes quando saímos para fora da casa abrimos a porta e saímos.
'''kgp-por-biltrans output: ''' ^Kejẽn<cir>/Às vezes<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^re<n>/fora<adv>/grama<n><f>/campo<n><m>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^kutẽ<v><iv><sg>/cair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ĩn<n>/casa<n><f>$ ^kãkã<v><tv>/cavoucar<vblex>/revistar<vblex>/vasculhar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jãnkã<n>/porta<n><f>$ ^nón<cir>/após<pr>/atrás<adv>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^króm ke<v><iv>/sair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: '''Às vezes #nós #ele, fora para #cair #tendo ido, #nós por #casa #cavoucar #sentar para, então #nós #ele, #porta após então #sair #tendo ido daí.
*'''Original sentence: ''' Kejẽn ẽg japry tóg jonjor sag tĩ, kỹ ẽg tóg ti rãgfãn tĩ, ẽg tỹ vẽnhrã hã ti tĩn jé.
'''Intended Portuguese translation: ''' Às vezes o nosso caminho tem muitas curvas, então nós o endireitamos para ficar mais reto.
'''kgp-por-biltrans output: ''' ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^japry<n>/caminho# de algo<n><m>/caminho# de alguém<n><m>$ ^tóg<su>/ele<prn><tn>$ ^jonjor<n>/curvas<n><f>/ziguezague<n><m>$ ^sag<v><tv>/atar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^rãgfãn<v><tv>/endireitar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^vẽnh rã hã<cir>/perto<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^tĩn<v><tv>/fazer andar<vblex>/passar por cima de<vblex>$ ^jé<n>/reza<n><f>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #nós #caminho# de algo #ele #curvas #atar #tendo ido, então #nós #ele #ele #endireitar #tendo ido, #nós por perto #ele #fazer andar #reza.
*'''Original sentence: ''' “Inh kã’u ã,” he tóg. “Inh mỹ ã tóg tỹ, ũ nĩ ven nĩ vẽ, hã ra ã tóg tỹ, ã hã nĩ”, he tóg.
'''Intended Portuguese translation: ''' “Você me assustou, pensei que você era um outro mas é você mesmo”, ele falou para mim.
'''kgp-por-biltrans output: ''' “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^kã’u<v><tv>/assustar<vblex>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$^,<cm>/,<cm>$” ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$ “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ven<v><tv><sg>/mostrar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^,<cm>/,<cm>$ ^hã<o>/igual<adj>/parecido<adj>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^hã<o>/igual<adj>/parecido<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$”^,<cm>/,<cm>$ ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' “#Eu #assustar #tu,” sim #ele. “#Eu para #tu #ele por, #alguém #sentar #mostrar #sentar #é, #igual para #tu #ele por, #tu #igual #sentar”, sim #ele.
*'''Original sentence: ''' Kyfé tỹ ag tóg, jagnẽ mĩ tỹ kririn ke ja nĩ.
'''Intended Portuguese translation: ''' Eles haviam se riscado um ao outro com uma faca.
'''kgp-por-biltrans output: ''' ^tỹ<cir>/por<pr>/com<pr>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jagnẽ<prn>/si mesmo<prn>$ ^mĩ<cir>/dentro<adv>$ ^tỹ<cir>/por<pr>/com<pr>$ ^kririn ke<v><iv>/riscar<vblex>$ ^ja<mod>/feito<adj>/terminado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Faca por #eles #ele, #si mesmo dentro por #riscar #feito #sentar.
*'''Original sentence: ''' Ẽg tỹ ũ mũmẽg han kỹ ẽg tóg, ti pãnĩ tá ti vóg tĩ; ẽg tỹ ti vatánh vẽ.
'''Intended Portuguese translation: ''' Quando queremos dar um susto em alguém, nós o pegamos por trás, assim nós o assustamos.
'''kgp-por-biltrans output: ''' ^Ẽg<prn><pes>/Nós<prn><tn><pes>/Nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ũ<prn>/alguém<prn><tn>$ ^mũmẽg<v><iv>/estar com medo<vblex>$ ^han<v><iv><sg>/sarar<vblex>/melhorar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vóg<v><tv>/mexer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^;<sent>/;<sent>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vatánh<v><tv>/enganar<vblex>/assustar<vblex>/contrariar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Nós por #alguém #estar com medo #sarar então #nós #ele, #ele atrás #ele #mexer #tendo ido; #nós por #ele #enganar #é.

==Additions==
*Added 103 words to the bilingual dictionary for "Polished RBMT System" assignment
==Final evaluation==
===kgp transducer===
*Precision against annotated.basic corpus:
*Recall against annotated.basic corpus:
*Coverage over the large corpus:
$ aq-covtest ling073-kgp-corpus/kgp.corpus.large.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 406588
Coverage: 91.99%
Top unknown words in the corpus:
726 nĩn
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
377 jafã
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
Translation time: 1.0561802387237549 seconds
*The number of words in the large corpus: 366,209
*The number of stems in the transducer: 2,299

===kgp-por pair===
*WER over longer corpus:
*PER over longer corpus:
*Proportion of stems translated correctly in the longer corpus:
*Trimmed coverage over longer and large corpora:
*The number of tokens in longer and large corpora:

[[Category:Sp19_TranslationPairs]] [[Category:Kaingang]] [[Category:Portuguese]]

Kaingang and Portuguese

2019-04-20T04:49:40Z

Drosset1: /* kgp transducer */

Resources for machine translation between [[Kaingang]] and [[Portuguese]]. [https://github.com/erick-guti-prado/apertium-por]

==External Resources==

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por Repository for language pair]

*[https://github.com/erick-guti-prado/apertium-por Portuguese transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp Kaingang transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por-corpus Parallel corpus]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Contrastive_Grammar Contrastive Grammar between Kaingang and Portuguese]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Lexical_Selection Lexical selection]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Structural_Transfer Structural Transfer]

==kgp → por Evaluation==
===Coverage (number of forms analysed) of Kaingang transducer on kgp.sentences.txt file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 32669
Coverage: 89.98%
Top unknown words in the corpus:
160 há
60 kaga
52 mẽg
47 sĩ
46 si
44 nĩn
43 kanẽ
42 Ka
40 ka
39 ké
37 fe
34 pẽn
33 rá
30 nĩm
27 vég
27 panh
26 jafã
26 isóg
24 jagy
24 Jesus
Translation time: 0.08736515045166016 seconds

===Coverage of Kaingang-Portuguese transducer (xyz-abc.automorf.hfst) on the same file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp-por/kgp-por.automorf.bin
Number of tokenised words in the corpus: 34281
Coverage: 58.82%
Top unknown words in the corpus:
739 vỹ
300 ki
287 fi
280 to
198 há
189 sóg
188 vẽnh
182 tũ
171 tá
145 nĩgtĩ
135 tĩg
134 ha
127 mẽ
121 pẽ
114 kar
114 mág
111 nén
111 kã
105 jẽ
102 kri
Translation time: 0.06977176666259766 seconds
===10 sentences===
*'''Original sentence: ''' Ag mỹ tóg, e tĩ.
'''Intended Portuguese translation: ''' Ficaram surpresos.
'''kgp-por-biltrans output: ''' ^Ag<prn><pes>/Eles<prn><tn><pes>/Deles<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Eles para #ele, #fazer #tendo ido.
*'''Original sentence: ''' Ti mré ke ag vẽ.
'''Intended Portuguese translation: ''' São os seus companheiros.
'''kgp-por-biltrans output: ''' ^Ti<prn><pes>/Ele<prn><tn><pes>/Lhe<prn><tn><pes>/Seu<prn><tn><pes>/Dele<prn><tn><pes>$ ^mré ke<mod>/parente<n><m>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Ele #parente #eles #é.
*'''Original sentence: ''' Kã’ég ũ tóg sá nĩ, ũ tóg kusũg nĩ, ũ tóg tánh nĩ, ũ tóg kógnár nĩ, e ti nỹtĩ.
'''Intended Portuguese translation: ''' Algumas araras são pretas, outras vermelhas, outras verdeazuis, outras pintadas, há muitas variedades.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^sá<mod>/preto<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kusũg<mod>/vermelho<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tánh<mod>/azul<adj>/verde<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kógnár<mod>/pintado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^nỹtĩ<v><iv><pl>/existir<vblex>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Arara #alguém #ele #preto #sentar, #alguém #ele #vermelho #sentar, #alguém #ele #azul #sentar, #alguém #ele #pintado #sentar, #fazer #ele #existir
*'''Original sentence: ''' Tãnh jũn kym ẽg tĩ, kỹ ẽg tóg, ti fár tar ẽn vãvãm tĩ, jó ẽg ti fár tãnãj ẽn mré ti kãtá ke ẽn ko tĩ, tãnh kaju ti.
'''Intended Portuguese translation: ''' Cortamos o palmito, jogamos fora a casca dura e comemos a casca mole com o coração.
'''kgp-por-biltrans output: ''' ^jũn<n>/palmito<n><m>$ ^kym<v><tv><sg>/cortar<vblex>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^*tar/*tar$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^vãvãm<v><tv><pl>/vender<vblex>/descartar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^jó<cir>/antes<adv>/em frente<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^tãnãj<n>/mole<adj><mf>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^mré<cir>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^*kãtá/*kãtá$ ^ke<n>/sobra<n><f>/resto<n><m>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^ko<v><tv>/usar<vblex>/comer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^tãnh<n>/palmeira<n><f>$ ^kaju<v>/empurrar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Palmeira #palmito #cortar #nós #tendo ido, então #nós #ele, #ele #pele *tar #aquele #vender #tendo ido, #antes #nós #ele #pele #mole #aquele com #ele *kãtá #sobra #aquele #usar #tendo ido, #palmeira #empurrar #ele.
*'''Original sentence: ''' Kejẽn ũ tóg ẽg pãnĩ tá ẽg jé prẽr tĩ, kỹ ẽg tóg, vĩrĩn ke kỹ ti mĩ ẽvãnh kỹ, ti mré vĩ mũ sir.
'''Intended Portuguese translation: ''' Às vezes alguém nos chama nas nossas costas, aí nós nos viramos para olhá-lo e falar com ele.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^jé prẽr<v><sv>/chamar<vblex><sv>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^vĩrĩn ke<v><iv>/virar-se<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mĩ<cir>/dentro<adv>$ ^ẽvãnh<v><iv>/enxergar<vblex>/observar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mré<cir>/com<pr>$ ^vĩ<n>/palavra<n><f>/discurso<n><m>$ ^mũ<a>/fazendo<vblex>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #alguém #ele #nós atrás #nós #chamar #tendo ido, então #nós #ele, #virar-se então #ele dentro #enxergar então, #ele com #palavra #fazendo daí.
*'''Original sentence: '''Kejẽn ẽg tóg, re ra kutẽ tĩ, ẽg tỹ ĩn kãkã nĩ ra, kỹ ẽg tóg, jãnkã nón kỹ króm ke tĩ sir.
'''Intended Portuguese translation: ''' Às vezes quando saímos para fora da casa abrimos a porta e saímos.
'''kgp-por-biltrans output: ''' ^Kejẽn<cir>/Às vezes<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^re<n>/fora<adv>/grama<n><f>/campo<n><m>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^kutẽ<v><iv><sg>/cair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ĩn<n>/casa<n><f>$ ^kãkã<v><tv>/cavoucar<vblex>/revistar<vblex>/vasculhar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jãnkã<n>/porta<n><f>$ ^nón<cir>/após<pr>/atrás<adv>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^króm ke<v><iv>/sair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: '''Às vezes #nós #ele, fora para #cair #tendo ido, #nós por #casa #cavoucar #sentar para, então #nós #ele, #porta após então #sair #tendo ido daí.
*'''Original sentence: ''' Kejẽn ẽg japry tóg jonjor sag tĩ, kỹ ẽg tóg ti rãgfãn tĩ, ẽg tỹ vẽnhrã hã ti tĩn jé.
'''Intended Portuguese translation: ''' Às vezes o nosso caminho tem muitas curvas, então nós o endireitamos para ficar mais reto.
'''kgp-por-biltrans output: ''' ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^japry<n>/caminho# de algo<n><m>/caminho# de alguém<n><m>$ ^tóg<su>/ele<prn><tn>$ ^jonjor<n>/curvas<n><f>/ziguezague<n><m>$ ^sag<v><tv>/atar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^rãgfãn<v><tv>/endireitar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^vẽnh rã hã<cir>/perto<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^tĩn<v><tv>/fazer andar<vblex>/passar por cima de<vblex>$ ^jé<n>/reza<n><f>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #nós #caminho# de algo #ele #curvas #atar #tendo ido, então #nós #ele #ele #endireitar #tendo ido, #nós por perto #ele #fazer andar #reza.
*'''Original sentence: ''' “Inh kã’u ã,” he tóg. “Inh mỹ ã tóg tỹ, ũ nĩ ven nĩ vẽ, hã ra ã tóg tỹ, ã hã nĩ”, he tóg.
'''Intended Portuguese translation: ''' “Você me assustou, pensei que você era um outro mas é você mesmo”, ele falou para mim.
'''kgp-por-biltrans output: ''' “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^kã’u<v><tv>/assustar<vblex>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$^,<cm>/,<cm>$” ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$ “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ven<v><tv><sg>/mostrar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^,<cm>/,<cm>$ ^hã<o>/igual<adj>/parecido<adj>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^hã<o>/igual<adj>/parecido<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$”^,<cm>/,<cm>$ ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' “#Eu #assustar #tu,” sim #ele. “#Eu para #tu #ele por, #alguém #sentar #mostrar #sentar #é, #igual para #tu #ele por, #tu #igual #sentar”, sim #ele.
*'''Original sentence: ''' Kyfé tỹ ag tóg, jagnẽ mĩ tỹ kririn ke ja nĩ.
'''Intended Portuguese translation: ''' Eles haviam se riscado um ao outro com uma faca.
'''kgp-por-biltrans output: ''' ^tỹ<cir>/por<pr>/com<pr>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jagnẽ<prn>/si mesmo<prn>$ ^mĩ<cir>/dentro<adv>$ ^tỹ<cir>/por<pr>/com<pr>$ ^kririn ke<v><iv>/riscar<vblex>$ ^ja<mod>/feito<adj>/terminado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Faca por #eles #ele, #si mesmo dentro por #riscar #feito #sentar.
*'''Original sentence: ''' Ẽg tỹ ũ mũmẽg han kỹ ẽg tóg, ti pãnĩ tá ti vóg tĩ; ẽg tỹ ti vatánh vẽ.
'''Intended Portuguese translation: ''' Quando queremos dar um susto em alguém, nós o pegamos por trás, assim nós o assustamos.
'''kgp-por-biltrans output: ''' ^Ẽg<prn><pes>/Nós<prn><tn><pes>/Nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ũ<prn>/alguém<prn><tn>$ ^mũmẽg<v><iv>/estar com medo<vblex>$ ^han<v><iv><sg>/sarar<vblex>/melhorar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vóg<v><tv>/mexer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^;<sent>/;<sent>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vatánh<v><tv>/enganar<vblex>/assustar<vblex>/contrariar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Nós por #alguém #estar com medo #sarar então #nós #ele, #ele atrás #ele #mexer #tendo ido; #nós por #ele #enganar #é.

==Additions==
*Added 103 words to the bilingual dictionary for "Polished RBMT System" assignment
==Final evaluation==
===kgp transducer===
*Precision against annotated.basic corpus:
*Recall against annotated.basic corpus:
*Coverage over the large corpus:
$ aq-covtest ling073-kgp-corpus/kgp.corpus.large.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 406588
Coverage: 91.99%
Top unknown words in the corpus:
726 nĩn
716 Cristo
609 vég
549 venh
480 nỹtĩn
470 Senhor
428 nĩm
392 hẽn
385 kuprĩg
377 jafã
375 cidade
350 henh
340 Paulo
334 jyvẽn
331 pir
308 mũꞌ
307 nĩꞌ
307 Pedro
289 régre
271 vin
Translation time: 1.0561802387237549 seconds
*The number of words in the large corpus:
*The number of stems in the transducer:

===kgp-por pair===
*WER over longer corpus:
*PER over longer corpus:
*Proportion of stems translated correctly in the longer corpus:
*Trimmed coverage over longer and large corpora:
*The number of tokens in longer and large corpora:

[[Category:Sp19_TranslationPairs]] [[Category:Kaingang]] [[Category:Portuguese]]

Kaingang and Portuguese

2019-04-20T04:32:08Z

Drosset1:

Resources for machine translation between [[Kaingang]] and [[Portuguese]]. [https://github.com/erick-guti-prado/apertium-por]

==External Resources==

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por Repository for language pair]

*[https://github.com/erick-guti-prado/apertium-por Portuguese transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp Kaingang transducer]

*[https://github.swarthmore.edu/Ling073-sp19/ling073-kgp-por-corpus Parallel corpus]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Contrastive_Grammar Contrastive Grammar between Kaingang and Portuguese]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Lexical_Selection Lexical selection]

*[https://wikis.swarthmore.edu/ling073/Kaingang_and_Portuguese/Structural_Transfer Structural Transfer]

==kgp → por Evaluation==
===Coverage (number of forms analysed) of Kaingang transducer on kgp.sentences.txt file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp/kgp.automorf.bin
Number of tokenised words in the corpus: 32669
Coverage: 89.98%
Top unknown words in the corpus:
160 há
60 kaga
52 mẽg
47 sĩ
46 si
44 nĩn
43 kanẽ
42 Ka
40 ka
39 ké
37 fe
34 pẽn
33 rá
30 nĩm
27 vég
27 panh
26 jafã
26 isóg
24 jagy
24 Jesus
Translation time: 0.08736515045166016 seconds

===Coverage of Kaingang-Portuguese transducer (xyz-abc.automorf.hfst) on the same file===
$ aq-covtest ling073-kgp-por-corpus/kgp.sentences.txt ling073-kgp-por/kgp-por.automorf.bin
Number of tokenised words in the corpus: 34281
Coverage: 58.82%
Top unknown words in the corpus:
739 vỹ
300 ki
287 fi
280 to
198 há
189 sóg
188 vẽnh
182 tũ
171 tá
145 nĩgtĩ
135 tĩg
134 ha
127 mẽ
121 pẽ
114 kar
114 mág
111 nén
111 kã
105 jẽ
102 kri
Translation time: 0.06977176666259766 seconds
===10 sentences===
*'''Original sentence: ''' Ag mỹ tóg, e tĩ.
'''Intended Portuguese translation: ''' Ficaram surpresos.
'''kgp-por-biltrans output: ''' ^Ag<prn><pes>/Eles<prn><tn><pes>/Deles<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Eles para #ele, #fazer #tendo ido.
*'''Original sentence: ''' Ti mré ke ag vẽ.
'''Intended Portuguese translation: ''' São os seus companheiros.
'''kgp-por-biltrans output: ''' ^Ti<prn><pes>/Ele<prn><tn><pes>/Lhe<prn><tn><pes>/Seu<prn><tn><pes>/Dele<prn><tn><pes>$ ^mré ke<mod>/parente<n><m>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Ele #parente #eles #é.
*'''Original sentence: ''' Kã’ég ũ tóg sá nĩ, ũ tóg kusũg nĩ, ũ tóg tánh nĩ, ũ tóg kógnár nĩ, e ti nỹtĩ.
'''Intended Portuguese translation: ''' Algumas araras são pretas, outras vermelhas, outras verdeazuis, outras pintadas, há muitas variedades.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^sá<mod>/preto<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kusũg<mod>/vermelho<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tánh<mod>/azul<adj>/verde<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^kógnár<mod>/pintado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^,<cm>/,<cm>$ ^e<v><iv>/fazer<vblex>/causar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^nỹtĩ<v><iv><pl>/existir<vblex>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Arara #alguém #ele #preto #sentar, #alguém #ele #vermelho #sentar, #alguém #ele #azul #sentar, #alguém #ele #pintado #sentar, #fazer #ele #existir
*'''Original sentence: ''' Tãnh jũn kym ẽg tĩ, kỹ ẽg tóg, ti fár tar ẽn vãvãm tĩ, jó ẽg ti fár tãnãj ẽn mré ti kãtá ke ẽn ko tĩ, tãnh kaju ti.
'''Intended Portuguese translation: ''' Cortamos o palmito, jogamos fora a casca dura e comemos a casca mole com o coração.
'''kgp-por-biltrans output: ''' ^jũn<n>/palmito<n><m>$ ^kym<v><tv><sg>/cortar<vblex>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^*tar/*tar$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^vãvãm<v><tv><pl>/vender<vblex>/descartar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^jó<cir>/antes<adv>/em frente<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^fár<n>/pele<n><f>/casca<n><f>$ ^tãnãj<n>/mole<adj><mf>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^mré<cir>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^*kãtá/*kãtá$ ^ke<n>/sobra<n><f>/resto<n><m>$ ^ẽn<prn><dem>/aquele<prn><tn><dem>$ ^ko<v><tv>/usar<vblex>/comer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^tãnh<n>/palmeira<n><f>$ ^kaju<v>/empurrar<vblex>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Palmeira #palmito #cortar #nós #tendo ido, então #nós #ele, #ele #pele *tar #aquele #vender #tendo ido, #antes #nós #ele #pele #mole #aquele com #ele *kãtá #sobra #aquele #usar #tendo ido, #palmeira #empurrar #ele.
*'''Original sentence: ''' Kejẽn ũ tóg ẽg pãnĩ tá ẽg jé prẽr tĩ, kỹ ẽg tóg, vĩrĩn ke kỹ ti mĩ ẽvãnh kỹ, ti mré vĩ mũ sir.
'''Intended Portuguese translation: ''' Às vezes alguém nos chama nas nossas costas, aí nós nos viramos para olhá-lo e falar com ele.
'''kgp-por-biltrans output: ''' ^ũ<prn>/alguém<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^jé prẽr<v><sv>/chamar<vblex><sv>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^vĩrĩn ke<v><iv>/virar-se<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mĩ<cir>/dentro<adv>$ ^ẽvãnh<v><iv>/enxergar<vblex>/observar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^mré<cir>/com<pr>$ ^vĩ<n>/palavra<n><f>/discurso<n><m>$ ^mũ<a>/fazendo<vblex>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #alguém #ele #nós atrás #nós #chamar #tendo ido, então #nós #ele, #virar-se então #ele dentro #enxergar então, #ele com #palavra #fazendo daí.
*'''Original sentence: '''Kejẽn ẽg tóg, re ra kutẽ tĩ, ẽg tỹ ĩn kãkã nĩ ra, kỹ ẽg tóg, jãnkã nón kỹ króm ke tĩ sir.
'''Intended Portuguese translation: ''' Às vezes quando saímos para fora da casa abrimos a porta e saímos.
'''kgp-por-biltrans output: ''' ^Kejẽn<cir>/Às vezes<adv>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^re<n>/fora<adv>/grama<n><f>/campo<n><m>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^kutẽ<v><iv><sg>/cair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ĩn<n>/casa<n><f>$ ^kãkã<v><tv>/cavoucar<vblex>/revistar<vblex>/vasculhar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jãnkã<n>/porta<n><f>$ ^nón<cir>/após<pr>/atrás<adv>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^króm ke<v><iv>/sair<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$ ^sir<o>/daí<cnjadv>/então<adv>/assim<adv>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: '''Às vezes #nós #ele, fora para #cair #tendo ido, #nós por #casa #cavoucar #sentar para, então #nós #ele, #porta após então #sair #tendo ido daí.
*'''Original sentence: ''' Kejẽn ẽg japry tóg jonjor sag tĩ, kỹ ẽg tóg ti rãgfãn tĩ, ẽg tỹ vẽnhrã hã ti tĩn jé.
'''Intended Portuguese translation: ''' Às vezes o nosso caminho tem muitas curvas, então nós o endireitamos para ficar mais reto.
'''kgp-por-biltrans output: ''' ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^japry<n>/caminho# de algo<n><m>/caminho# de alguém<n><m>$ ^tóg<su>/ele<prn><tn>$ ^jonjor<n>/curvas<n><f>/ziguezague<n><m>$ ^sag<v><tv>/atar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^rãgfãn<v><tv>/endireitar<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^,<cm>/,<cm>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^vẽnh rã hã<cir>/perto<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^tĩn<v><tv>/fazer andar<vblex>/passar por cima de<vblex>$ ^jé<n>/reza<n><f>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' Às vezes #nós #caminho# de algo #ele #curvas #atar #tendo ido, então #nós #ele #ele #endireitar #tendo ido, #nós por perto #ele #fazer andar #reza.
*'''Original sentence: ''' “Inh kã’u ã,” he tóg. “Inh mỹ ã tóg tỹ, ũ nĩ ven nĩ vẽ, hã ra ã tóg tỹ, ã hã nĩ”, he tóg.
'''Intended Portuguese translation: ''' “Você me assustou, pensei que você era um outro mas é você mesmo”, ele falou para mim.
'''kgp-por-biltrans output: ''' “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^kã’u<v><tv>/assustar<vblex>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$^,<cm>/,<cm>$” ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$ “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^ven<v><tv><sg>/mostrar<vblex>$ ^nĩ<v><iv><sg>/sentar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^,<cm>/,<cm>$ ^hã<o>/igual<adj>/parecido<adj>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ã<prn>/tu<prn><tn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/de ti<prn><tn>/de você<prn><tn>$ ^hã<o>/igual<adj>/parecido<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$”^,<cm>/,<cm>$ ^he<intj>/sim<adv>/aceito<adj>/vou fazer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' “#Eu #assustar #tu,” sim #ele. “#Eu para #tu #ele por, #alguém #sentar #mostrar #sentar #é, #igual para #tu #ele por, #tu #igual #sentar”, sim #ele.
*'''Original sentence: ''' Kyfé tỹ ag tóg, jagnẽ mĩ tỹ kririn ke ja nĩ.
'''Intended Portuguese translation: ''' Eles haviam se riscado um ao outro com uma faca.
'''kgp-por-biltrans output: ''' ^tỹ<cir>/por<pr>/com<pr>$ ^ag<prn><pes>/eles<prn><tn><pes>/deles<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^jagnẽ<prn>/si mesmo<prn>$ ^mĩ<cir>/dentro<adv>$ ^tỹ<cir>/por<pr>/com<pr>$ ^kririn ke<v><iv>/riscar<vblex>$ ^ja<mod>/feito<adj>/terminado<adj>$ ^nĩ<v><iv><sg>/sentar<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Faca por #eles #ele, #si mesmo dentro por #riscar #feito #sentar.
*'''Original sentence: ''' Ẽg tỹ ũ mũmẽg han kỹ ẽg tóg, ti pãnĩ tá ti vóg tĩ; ẽg tỹ ti vatánh vẽ.
'''Intended Portuguese translation: ''' Quando queremos dar um susto em alguém, nós o pegamos por trás, assim nós o assustamos.
'''kgp-por-biltrans output: ''' ^Ẽg<prn><pes>/Nós<prn><tn><pes>/Nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ũ<prn>/alguém<prn><tn>$ ^mũmẽg<v><iv>/estar com medo<vblex>$ ^han<v><iv><sg>/sarar<vblex>/melhorar<vblex>$ ^kỹ<conj>/então<adv>/por causa de<n><f>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tóg<su>/ele<prn><tn>$^,<cm>/,<cm>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^pãnĩ tá<cir>/atrás<adv>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vóg<v><tv>/mexer<vblex>$ ^tĩ<a>/tendo ido<vblex>/habitualmente<adv>$^;<sent>/;<sent>$ ^ẽg<prn><pes>/nós<prn><tn><pes>/nosso<prn><tn><pes>$ ^tỹ<cir>/por<pr>/com<pr>$ ^ti<prn><pes>/ele<prn><tn><pes>/lhe<prn><tn><pes>/seu<prn><tn><pes>/dele<prn><tn><pes>$ ^vatánh<v><tv>/enganar<vblex>/assustar<vblex>/contrariar<vblex>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^.<sent>/.<sent>$^.<sent>/.<sent>$
'''kgp-por output: ''' #Nós por #alguém #estar com medo #sarar então #nós #ele, #ele atrás #ele #mexer #tendo ido; #nós por #ele #enganar #é.

==Additions==
*Added 103 words to the bilingual dictionary for "Polished RBMT System" assignment
==Final evaluation==
===kgp transducer===
*Precision against annotated.basic corpus:
*Recall against annotated.basic corpus:
*Coverage over the large corpus:
*The number of words in the large corpus:
*The number of stems in the transducer:
===kgp-por pair===
*WER over longer corpus:
*PER over longer corpus:
*Proportion of stems translated correctly in the longer corpus:
*Trimmed coverage over longer and large corpora:
*The number of tokens in longer and large corpora:

[[Category:Sp19_TranslationPairs]] [[Category:Kaingang]] [[Category:Portuguese]]

Kaingang and Portuguese/Structural Transfer

2019-04-19T19:16:17Z

Drosset1: /* Example for implementation */

[[Category:Sp19_StructuralTransfer]] [[Category:Kaingang]] [[Category:Portuguese]]
==Pre-evaluation==
Statistics about input files
-------------------------------------------------------
Number of words in reference: 63
Number of words in test: 63
Number of unknown words (marked with a star) in test: 26
Percentage of unknown words: 41.27 %

Results when removing unknown-word marks (stars)
-------------------------------------------------------
Edit distance: 38
Word error rate (WER): 60.32 %
Number of position-independent correct words: 25
Position-independent word error rate (PER): 60.32 %

Results when unknown-word marks (stars) are not removed
-------------------------------------------------------
Edit distance: 63
Word Error Rate (WER): 100.00 %
Number of position-independent correct words: 0
Position-independent word error rate (PER): 100.00 %

Statistics about the translation of unknown words
-------------------------------------------------------
Number of unknown words which were free rides: 25
Percentage of unknown words that were free rides: 96.15 %

*'''WER:''' 60.32%

*'''PER:''' 60.32%

*'''Coverage:'''
$ aq-covtest ling073-kgp-por-corpus/kgp.tests.txt ling073-kgp-por/kgp-por.automorf.bin
Number of tokenised words in the corpus: 82
Coverage: 64.63%
Top unknown words in the corpus:
3 vỹ
2 fi
1 kafã
1 Nũgnũj
1 kur
1 ẽgno
1 tũg
1 São
1 Pau
1 o
1 rã
1 jur
1 tá
1 Téj
1 ki
1 panh
1 kãfór
1 kyrũ
1 jãmré
1 ũn
Translation time: 0.0028295516967773438 seconds
==Example for implementation==
'''Sentence:'''
“Inh kã’u ã,” he tóg. “Inh mỹ ã tóg tỹ, ũ nĩ ven nĩ vẽ, hã ra ã tóg tỹ, ã hã nĩ”, he tóg.

'''Portuguese Translation:'''
“Você me assustou, pensei que você era um outro mas é você mesmo”, ele falou para mim.

'''Tagger output:'''
“^Inh/inh<prn><pes>/inh<prn><p1><sg>$ ^kã’u/kã’u<v><tv>$ ^ã/ã<prn>/ã<prn><p2><sg>$^,/,<cm>$” ^he/he<v><tv>$ ^tóg/tóg<su>$^./.<sent>$ “^Inh/inh<prn><pes>/inh<prn><p1><sg>$ ^mỹ/mỹ<cir>$ ^ã/ã<prn>/ã<prn><p2><sg>$ ^tóg/tóg<su>$ ^tỹ/tỹ<cir>$^,/,<cm>$ ^ũ/ũ<prn>$ ^nĩ/nĩ<a>/nĩ<n>/nĩ<v><iv><sg>$ ^ven/ven<v><tv><sg>$ ^nĩ/nĩ<a>/nĩ<n>/nĩ<v><iv><sg>$ ^vẽ/vẽ<a>$^,/,<cm>$ ^hã/hã<o>$ ^ra/ra<cir>$ ^ã/ã<prn>/ã<prn><p2><sg>$ ^tóg/tóg<su>$ ^tỹ/tỹ<cir>$^,/,<cm>$ ^ã/ã<prn>/ã<prn><p2><sg>$ ^hã/hã<o>$ ^nĩ/nĩ<a>/nĩ<n>/nĩ<v><iv><sg>$”^,/,<cm>$ ^he/he<v><tv>$ ^tóg/tóg<su>$^./.<sent>$^./.<sent>$

'''Biltrans output:'''
“^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^kã’u<v><tv>/assustar<vblex>$ ^ã<prn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/tu<prn><tn><p2><mf><sg>$^,<cm>/,<cm>$” ^he<v><tv>/dizer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$ “^Inh<prn><pes>/Eu<prn><tn><pes>/Meu<prn><tn><pes>/De mim<prn><tn><pes>$ ^mỹ<cir>/para<pr>$ ^ã<prn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/tu<prn><tn><p2><mf><sg>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ũ<prn>/alguém<prn><tn>$ ^nĩ<a>/no momento<adv>$ ^ven<v><tv><sg>/mostrar<vblex>$ ^nĩ<a>/no momento<adv>$ ^vẽ<a>/é<vblex>/era<vblex>/ser<vblex>$^,<cm>/,<cm>$ ^hã<o>/igual<adj>/parecido<adj>$ ^ra<cir>/para<pr>/apesar do<cnjadv>$ ^ã<prn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/tu<prn><tn><p2><mf><sg>$ ^tóg<su>/ele<prn><tn>$ ^tỹ<cir>/por<pr>/com<pr>$^,<cm>/,<cm>$ ^ã<prn>/teu<prn><tn>/seu<prn><tn>/você<prn><tn>/tu<prn><tn><p2><mf><sg>$ ^hã<o>/igual<adj>/parecido<adj>$ ^nĩ<a>/no momento<adv>$”^,<cm>/,<cm>$ ^he<v><tv>/dizer<vblex>$ ^tóg<su>/ele<prn><tn>$^.<sent>/.<sent>$^.<sent>/.<sent>

'''Chunker output:'''
apertium-transfer: Rule 1 .<sent>/.<sent>
apertium-transfer: Rule 2 mỹ<cir>/para<pr>
apertium-transfer: Rule 3 ã<prn>/tu<prn><tn><p2><mf><sg> tóg<su>/ele<prn><tn>
apertium-transfer: Rule 2 tỹ<cir>/por<pr>/com<pr>
apertium-transfer: Rule 2 ra<cir>/para<pr>/apesar do<cnjadv>
apertium-transfer: Rule 3 ã<prn>/tu<prn><tn><p2><mf><sg> tóg<su>/ele<prn><tn>
apertium-transfer: Rule 2 tỹ<cir>/por<pr>/com<pr>
apertium-transfer: Rule 1 .<sent>/.<sent>
apertium-transfer: Rule 1 .<sent>/.<sent>
“^default<default>{^Eu<prn><tn><pes>$}$ ^default<default>{^assustar<vblex>$}$ ^default<default>{^tu<prn><tn><p2><mf><sg>$}$^default<default>{^,<cm>$}$” ^default<default>{^dizer<vblex>$}$ ^default<default>{^ele<prn><tn>$}$^sent<SENT>{^.<sent>$}$ “^default<default>{^Eu<prn><tn><pes>$}$ ^pr<SP>{^para<pr>$}$ ^prn<SN><CD>{^tu<prn><tn><p2><mf><sg><2>$}$ ^pr<SP>{^por<pr>$}$^default<default>{^,<cm>$}$ ^default<default>{^alguém<prn><tn>$}$ ^default<default>{^no momento<adv>$}$ ^default<default>{^mostrar<vblex>$}$ ^default<default>{^no momento<adv>$}$ ^default<default>{^é<vblex>$}$^default<default>{^,<cm>$}$ ^default<default>{^igual<adj>$}$ ^pr<SP>{^para<pr>$}$ ^prn<SN><CD>{^tu<prn><tn><p2><mf><sg><2>$}$ ^pr<SP>{^por<pr>$}$^default<default>{^,<cm>$}$ ^default<default>{^tu<prn><tn><p2><mf><sg>$}$ ^default<default>{^igual<adj>$}$ ^default<default>{^no momento<adv>$}$”^default<default>{^,<cm>$}$ ^default<default>{^dizer<vblex>$}$ ^default<default>{^ele<prn><tn>$}$^sent<SENT>{^.<sent>$}$^sent<SENT>{^.<sent>$}$

'''Interchunk output:'''
apertium-interchunk: Rule 1 prn<SN><CD>{^tu<prn><tn><p2><mf><sg><2>$}
apertium-interchunk: Rule 2 prn<SN><CD>{^tu<prn><tn><p2><mf><sg><2>$} pr<SP>{^por<pr>$}
apertium-interchunk: Rule 1 prn<SN><CD>{^tu<prn><tn><p2><mf><sg><2>$}
apertium-interchunk: Rule 2 prn<SN><CD>{^tu<prn><tn><p2><mf><sg><2>$} pr<SP>{^por<pr>$}
“^default<default>{^Eu<prn><tn><pes>$}$ ^default<default>{^assustar<vblex>$}$ ^default<default>{^tu<prn><tn><p2><mf><sg>$}$^default<default>{^,<cm>$}$” ^default<default>{^dizer<vblex>$}$ ^default<default>{^ele<prn><tn>$}$^sent<SENT>{^.<sent>$}$ “^default<default>{^Eu<prn><tn><pes>$}$ ^pr<SP>{^para<pr>$}$ ^pr<SP>{^por<pr>$}$ ^prn<SN><obj>{^tu<prn><tn><p2><mf><sg><2>$}$^default<default>{^,<cm>$}$ ^default<default>{^alguém<prn><tn>$}$ ^default<default>{^no momento<adv>$}$ ^default<default>{^mostrar<vblex>$}$ ^default<default>{^no momento<adv>$}$ ^default<default>{^é<vblex>$}$^default<default>{^,<cm>$}$ ^default<default>{^igual<adj>$}$ ^pr<SP>{^para<pr>$}$ ^pr<SP>{^por<pr>$}$ ^prn<SN><obj>{^tu<prn><tn><p2><mf><sg><2>$}$^default<default>{^,<cm>$}$ ^default<default>{^tu<prn><tn><p2><mf><sg>$}$ ^default<default>{^igual<adj>$}$ ^default<default>{^no momento<adv>$}$”^default<default>{^,<cm>$}$ ^default<default>{^dizer<vblex>$}$ ^default<default>{^ele<prn><tn>$}$^sent<SENT>{^.<sent>$}$^sent<SENT>{^.<sent>$}$

'''Postchunk output:'''
“^Eu<prn><tn><pes>$ ^assustar<vblex>$ ^tu<prn><tn><p2><mf><sg>$^,<cm>$” ^dizer<vblex>$ ^ele<prn><tn>$^.<sent>$ “^Eu<prn><tn><pes>$ ^para<pr>$ ^por<pr>$ ^tu<prn><tn><p2><mf><sg><obj>$^,<cm>$ ^alguém<prn><tn>$ ^no momento<adv>$ ^mostrar<vblex>$ ^no momento<adv>$ ^é<vblex>$^,<cm>$ ^igual<adj>$ ^para<pr>$ ^por<pr>$ ^tu<prn><tn><p2><mf><sg><obj>$^,<cm>$ ^tu<prn><tn><p2><mf><sg>$ ^igual<adj>$ ^no momento<adv>$”^,<cm>$ ^dizer<vblex>$ ^ele<prn><tn>$^.<sent>$^.<sent>$

'''kgp-por output:'''
“#Eu #assustar #tu,” #dizer #ele. “#Eu para por ti, #alguém #no momento #mostrar #no momento #é, #igual para por ti, #tu #igual #no momento”, #dizer #ele.

==Post-evaluation==
$ apertium-eval-translator -r ../tools/kgp.tests.txt -t ../ling073-kgp-por/kgp-por.tests.txt
Test file: '../ling073-kgp-por/kgp-por.tests.txt'
Reference file '../tools/kgp.tests.txt'

Statistics about input files
-------------------------------------------------------
Number of words in reference: 63
Number of words in test: 73
Number of unknown words (marked with a star) in test:
Percentage of unknown words: 0.00 %

Results when removing unknown-word marks (stars)
-------------------------------------------------------
Edit distance: 72
Word error rate (WER): 114.29 %
Number of position-independent correct words: 1
Position-independent word error rate (PER): 114.29 %

Results when unknown-word marks (stars) are not removed
-------------------------------------------------------
Edit distance: 72
Word Error Rate (WER): 114.29 %
Number of position-independent correct words: 1
Position-independent word error rate (PER): 114.29 %

Statistics about the translation of unknown words
-------------------------------------------------------
Number of unknown words which were free rides: 0
Percentage of unknown words that were free rides: 0%
*'''WER:''' 114.29 % ???
*'''PER:''' 114.29 % ???
*'''Coverage:'''
$ aq-covtest ling073-kgp-por-corpus/kgp.tests.txt ling073-kgp-por/kgp-por.automorf.bin
Number of tokenised words in the corpus: 80
Coverage: 100.00%
Top unknown words in the corpus:
Translation time: 0.002888202667236328 seconds

Kaingang and Portuguese/Structural Transfer

2019-04-19T19:15:44Z

Drosset1: