Difference between revisions of "Urum and English"

From LING073
Jump to: navigation, search
(→‎Resources developed for Urum)
(→‎Proportion of Stems Translated Correctly in Longer Corpus)
 
(47 intermediate revisions by 2 users not shown)
Line 3: Line 3:
 
* [https://wikis.swarthmore.edu/ling073/English LING073 English Language Documentation]
 
* [https://wikis.swarthmore.edu/ling073/English LING073 English Language Documentation]
  
−
== Resources Ceveloped for Urum ==
+
* [https://wikis.swarthmore.edu/ling073/index.php?title=Urum_and_English Urum and English Translation Pairs] ([https://github.swarthmore.edu/Ling073-sp23/ling073-uum-eng.git Source Code])
 +
** [https://wikis.swarthmore.edu/ling073/Urum_and_English/Lexical_Selection LING073 Urum-to-English Lexical Selection Examples]
 +
** [https://wikis.swarthmore.edu/ling073/Urum_and_English/Contrastive_Grammar LING073 Urum-to-English Conrastive Grammar Examples]
 +
** [https://wikis.swarthmore.edu/ling073/Urum_and_English/Structural_Transfer LING073 Urum-to-English Structural Transfer Examples]
 +
 
 +
== Resources Developed for Urum ==
 +
 
 
* [https://wikis.swarthmore.edu/ling073/Urum/Keyboard Urum Keyboard Layout] ([https://github.swarthmore.edu/Ling073-sp23/ling073-uum-keyboard Source Code])
 
* [https://wikis.swarthmore.edu/ling073/Urum/Keyboard Urum Keyboard Layout] ([https://github.swarthmore.edu/Ling073-sp23/ling073-uum-keyboard Source Code])
 +
* [https://wikis.swarthmore.edu/ling073/Urum/Grammar Urum Morphological Analyser] ([https://github.swarthmore.edu/Ling073-sp23/ling073-uum.git Source Code])
 
* [https://github.swarthmore.edu/Ling073-sp23/ling073-uum-tools Urum Language Tools Respository]
 
* [https://github.swarthmore.edu/Ling073-sp23/ling073-uum-tools Urum Language Tools Respository]
−
* [https://wikis.swarthmore.edu/ling073/Urum/Grammar Urum Morphological Analyser] ([https://github.swarthmore.edu/Ling073-sp23/ling073-uum.git Source Code])
 
  
 
== External resources ==
 
== External resources ==
Line 17: Line 23:
 
* [https://wikis.swarthmore.edu/ling073/index.php?title=Urum_and_English Urum and English Translation Pairs] ([https://github.swarthmore.edu/Ling073-sp23/ling073-uum-eng.git Source Code])
 
* [https://wikis.swarthmore.edu/ling073/index.php?title=Urum_and_English Urum and English Translation Pairs] ([https://github.swarthmore.edu/Ling073-sp23/ling073-uum-eng.git Source Code])
  
−
== uum → eng evaluation ==
+
== Transducer Coverage Over Example Text ==
−
COVERAGE
+
'''Urum Transducer Coverage:'''
 +
coverage: 155 / 222 (~0.69819819819819819820)
 +
remaining unknown forms: 67
 +
 
 +
'''Bilingual Transducer Coverage:'''
 +
coverage: 130 / 206 (~0.63106796116504854369)
 +
remaining unknown forms: 76
 +
 
 +
== (UUM ‣ ENG) Translation Evaluation ==
 +
 
 +
=== Urum Text ===
 +
 
 +
"Башла", - дэдим - башламадым,
 +
башлавужу йылдыз доғду.
 +
Панайана Вифлейемдэ тэна йер булмады кӧйд`е.
 +
Йосиф илен т`етти чӧльд`е, пичен илен йасла булду.
 +
Вахыт т`ельди - Христос доғду, сарып йасла чине хойду.
 +
Кӱн доғуштан ышхын доғду,
 +
ышихындан йылдыз кӧрӱндӱ,
 +
чобанлар да хорхундулар:
 +
"Бу нас йарых бу?"- дэдилер.
 +
Кӧктэн мелайиклер энди, чобанлара гӧңӱль верди:
 +
"Бӱйӱк аллах доғду, - дэди, -,
 +
варың да баш уруң" ,- дэди.
 +
Мелайиклер хонуштулар, азиз тӱркӱ йырладылар.
 +
Кӱн доғуштан йылдыз кельди, айдавҗусу бӱльбӱль эди.
 +
Учхун йелир, йундуз хонду, чобаллара мельэм олду, -
 +
Вифлейемдэ Христос доғду.
 +
Ол бешиге эништилер, ичиндэ Христос кӧрдӱлер, севиндилер доғушуна, бу дӱнйайа келишине.
 +
Бӱйӱгӱ - оҗа кӧрдӱ,
 +
ортанҗысы - игит кӧрдӱ,
 +
эң кичиги - тахтын булду,
 +
башына пуллар кийдирди.
 +
Не ахыллы халх эдилер,
 +
кӧктэн, йылдыздан бильдилер,
 +
мирон, алтын т`етирдилер, хозу бағыш т`етирдилер.
 +
Ирод падышах ишитти, аст`ер алып йола чыхты, йӱз он дӧрт бин бала хырды, Христозу да буламады.
 +
Аллах оғлуну сахлады,
 +
Ироднуң йолун бағлады.
 +
Христос т`етти Йегипитке, онда ӧстӱ, бӱмӱк олду, җеннет хапӱсӱн ачтырды, халхны йӱнахтан хутхарды.
 +
 
 +
=== English Translation ===
 +
 
 +
"Start!" I said but did not start,
 +
the dawn closed.
 +
The Holy Mary in Bethlehem did not find a peaceful place in the village.
 +
she went with Joseph to the field, found a manger with hay.
 +
The time has come - Christ is born, wrapped and laid in a manger.
 +
In the east, your light shone,
 +
a star appeared in the light,
 +
the shepherds were frightened:
 +
"What is this light?" they said.
 +
Angels descended from heavens, calmed the shepherds:
 +
"A great God is born," they said,
 +
"go and worship," they said.
 +
Angels proclaimed, sang holy songs.
 +
A star came from the east, accompanied by a nightingale.
 +
The light approached, the star descended, the shepherds were comforted -
 +
Christ was born in Bethlehem.
 +
They descended to that crib, saw Christ in it, rejoiced at his birth, his coming into this world.
 +
The eldest saw a husband,
 +
the middle saw a knight,
 +
the youngest found his throne,
 +
they put glitter on his head.
 +
What wise people they were -
 +
they learned from the star in the sky,
 +
they brought myrrh, gold, and a lamb as a gift.
 +
King Herod searched, set out on the road with a star, killed fourteen thousand children, but couldn't find Christ.
 +
Allah protected His son,
 +
King Herod heard.
 +
Christ went to Egypt, There He grew up and became great, opened the doors of paradise, saved people from sin.
 +
 
 +
=== Output of uum-eng-biltrans (Lexical Transfer) Mode ===
 +
"^Башла<v><tv><imp><p1><sg>/Start<vblex><tv><imp><p1><sg>$", - ^дэ<v><iv><past><p1><sg>/say<vblex><iv><past><p1><sg>$ - ^башла<v><tv><neg><past><p1><sg>/start<vblex><tv><neg><past><p1><sg>$,
 +
^башлавужу<n><nom>/founder<n><nom>$ ^йылдыз<n><nom>/star<n><nom>$ ^доғ<v><iv><past><p3><sg>/be# born<vblex><past><p3><sg>$.
 +
^Панайана<np><ant><f><sg><nom>/Mary<np><ant><f><sg><nom>$ ^Вифлейем<np><top><sg><loc>/Bethlehem<np><top><sg><loc>$ ^тэна<adj>/peaceful<adj>$ ^йер<n><nom>/place<n><nom>$ ^бул<v><tv><neg><past><p3><sg>/find<vblex><tv><neg><past><p3><sg>$ ^кӧй<n><loc>/village<n><loc>$.
 +
^Йосиф<np><ant><m><sg><nom>/Joseph<np><ant><m><sg><nom>$ ^илен<pr>/with<pr>$ ^тʼет<v><iv><past><p3><sg>/go<vblex><iv><past><p3><sg>$ ^чӧль<n><loc>/field<n><loc>$, ^пичен<n><nom>/hay<n><nom>$ ^илен<pr>/with<pr>$ ^йасла<n><nom>/manger<n><nom>$ ^бул<v><tv><past><p3><sg>/find<vblex><tv><past><p3><sg>$.
 +
^Вахыт<n><nom>/Time<n><nom>$ ^тʼель<v><iv><past><p3><sg>/come<vblex><iv><past><p3><sg>$ - ^Христос<np><ant><m><sg><nom>/Christ<np><ant><m><sg><nom>$ ^доғ<v><iv><past><p3><sg>/be# born<vblex><past><p3><sg>$, ^*сарып/*сарып$ ^йасла<n><nom>/manger<n><nom>$ ^чине<pr>/in<pr>$ ^хой<v><tv><past><p3><sg>/put<vblex><tv><past><p3><sg>$.
 +
^Кӱн<n><nom>/Sun<n><nom>/Day<n><nom>$ ^доғуш<n><abl>/birth<n><abl>$ ^ышхын<n><nom>/light<n><nom>$ ^доғ<v><iv><past><p3><sg>/be# born<vblex><past><p3><sg>$,
 +
^ышхын<n><abl>/light<n><abl>$ ^йылдыз<n><nom>/star<n><nom>$ ^кӧрӱн<v><iv><past><p3><sg>/emerge<vblex><iv><past><p3><sg>$,
 +
^чобан<n><pl><nom>/shepherd<n><pl><nom>$ ^да<part>/$ ^хорхун<v><iv><past><p3><pl>/frighten<vblex><past><p3><pl>$:
 +
"^Бу<det>/This<det><dem><sg>$ ^нас<adv>/how<adv><itg>$ ^йарых<adj>/light<adj><sint>$ ^бу<det>/this<det><dem><sg>$?"- ^дэ<v><iv><past><p3><pl>/say<vblex><iv><past><p3><pl>$.
 +
^Кӧк<n><abl>/Sky<n><abl>$ ^мелайик<n><pl><nom>/angel<n><pl><nom>$ ^эн<v><iv><past><p3><sg>/descend<vblex><iv><past><p3><sg>$, ^чобан<n><pl><dat>/shepherd<n><pl><dat>$ ^гӧңӱль<n><nom>/relief<n><nom>$ ^вер<v><tv><past><p3><sg>/bring<vblex><tv><past><p3><sg>$:
 +
"^Бӱйӱк<adj>/Big<adj><sint>$ ^аллах<n><nom>/God<n><nom>$ ^доғ<v><iv><past><p3><sg>/be# born<vblex><past><p3><sg>$, - ^дэ<v><iv><past><p3><sg>/say<vblex><iv><past><p3><sg>$, -,
 +
^*варың/*варың$ ^да<part>/$ ^баш<n><nom>/head<n><nom>$ ^уруң<v><iv><imp><p1><sg>/bow<vblex><iv><imp><p1><sg>$" ,- ^дэ<v><iv><past><p3><sg>/say<vblex><iv><past><p3><sg>$.
 +
^Мелайик<n><pl><nom>/Angel<n><pl><nom>$ ^хонуш<v><iv><past><p3><pl>/bless<vblex><iv><past><p3><pl>$, ^азиз<adj>/holy<adj><sint>$ ^тӱрк<n><acc>/song<n><acc>$ ^йырла<v><iv><past><p3><pl>/sing<vblex><iv><past><p3><pl>/йырла<vblex><iv><past><p3><pl>$.
 +
^Кӱн<n><nom>/Sun<n><nom>/Day<n><nom>$ ^доғуш<n><abl>/birth<n><abl>$ ^йылдыз<n><nom>/star<n><nom>$ ^кель<v><iv><past><p3><sg>/come<vblex><iv><past><p3><sg>$, ^*айдавҗусу/*айдавҗусу$ ^бӱльбӱль<n><nom>/nightingale<n><nom>$ ^э<v><tv><past><p3><sg>/be<vblex><tv><past><p3><sg>$.
 +
^Учхун<n><nom>/Spark<n><nom>$ ^*йелир/*йелир$, ^йундуз<n><nom>/star<n><nom>$ ^хон<v><iv><past><p3><sg>/fall<vblex><iv><past><p3><sg>$, ^*чобаллара/*чобаллара$ ^*мельэм/*мельэм$ ^ол<v><iv><past><p3><sg>/be<vblex><iv><past><p3><sg>/be able<vblex><iv><past><p3><sg>$, -
 +
^Вифлейем<np><top><sg><loc>/Bethlehem<np><top><sg><loc>$ ^Христос<np><ant><m><sg><nom>/Christ<np><ant><m><sg><nom>$ ^доғ<v><iv><past><p3><sg>/be# born<vblex><past><p3><sg>$.
 +
^Ол<v><iv><imp><p1><sg>/Be<vblex><iv><imp><p1><sg>/Be able<vblex><iv><imp><p1><sg>$ ^*бешиге/*бешиге$ ^*эништилер/*эништилер$, ^ич<n><px3sg><loc>/stomach<n><px3sg><loc>/interior<n><px3sg><loc>$ ^Христос<np><ant><m><sg><nom>/Christ<np><ant><m><sg><nom>$ ^кӧр<v><tv><past><p3><pl>/see<vblex><tv><past><p3><pl>$, ^*севиндилер/*севиндилер$ ^доғуш<n><px3sg><dat>/birth<n><px3sg><dat>$, ^бу<det>/this<det><dem><sg>$ ^*дӱнйайа/*дӱнйайа$ ^*келишине/*келишине$.
 +
^*Бӱйӱгӱ/*Бӱйӱгӱ$ - ^*оҗа/*оҗа$ ^кӧр<v><tv><past><p3><sg>/see<vblex><tv><past><p3><sg>$,
 +
^*ортанҗысы/*ортанҗысы$ - ^*игит/*игит$ ^кӧр<v><tv><past><p3><sg>/see<vblex><tv><past><p3><sg>$,
 +
^*эң/*эң$ ^*кичиги/*кичиги$ - ^*тахтын/*тахтын$ ^бул<v><tv><past><p3><sg>/find<vblex><tv><past><p3><sg>$,
 +
  ^баш<n><px3sg><dat>/head<n><px3sg><dat>$ ^*пуллар/*пуллар$ ^*кийдирди/*кийдирди$.
 +
^Не<prn><det><dem>/What<prn><det><dem>$ ^*ахыллы/*ахыллы$ ^*халх/*халх$ ^э<v><tv><past><p3><pl>/be<vblex><tv><past><p3><pl>$,
 +
^кӧк<n><abl>/sky<n><abl>$, ^йылдыз<n><abl>/star<n><abl>$ ^*бильдилер/*бильдилер$,
 +
^мирон<n><nom>/myrrh<n><nom>$, ^алтын<n><nom>/gold<n><nom>$ ^тʼетир<v><tv><past><p3><pl>/bring<vblex><tv><past><p3><pl>$, ^хоз<n><acc>/lamb<n><acc>$ ^бағыш<n><nom>/gift<n><nom>$ ^тʼетир<v><tv><past><p3><pl>/bring<vblex><tv><past><p3><pl>$.
 +
^Ирод<np><ant><m><sg><nom>/Herod<np><ant><m><sg><nom>$ ^падышах<n><nom>/king<n><nom>$ ^*ишитти/*ишитти$, ^астʼер<n><nom>/army<n><nom>$ ^*алып/*алып$ ^йол<n><dat>/path<n><dat>$ ^чых<v><iv><past><p3><sg>/go<vblex><iv><past><p3><sg>/come<vblex><iv><past><p3><sg>$, ^йӱз<num>/hundred<num><pl>$ ^он дӧрт<num>/fourteen<num><pl>$ ^бин<num>/thousand<num><pl>$ ^бала<n><nom>/child<n><nom>$ ^хыр<v><tv><past><p3><sg>/kill<vblex><tv><past><p3><sg>$, ^*Христозу/*Христозу$ ^да<part>/$ ^була<v><tv><neg><past><p3><sg>/find<vblex><tv><neg><past><p3><sg>$.
 +
^Аллах<n><nom>/God<n><nom>$ ^*оғлуну/*оғлуну$ ^*сахлады/*сахлады$,
 +
^*Ироднуң/*Ироднуң$ ^йол<n><px3sg><nom>/path<n><px3sg><nom>$ ^бағла<v><tv><past><p3><sg>/tie<vblex><tv><past><p3><sg>$.
 +
^Христос<np><ant><m><sg><nom>/Christ<np><ant><m><sg><nom>$ ^тʼет<v><iv><past><p3><sg>/go<vblex><iv><past><p3><sg>$ ^*Йегипитке/*Йегипитке$, ^онда<adv>/there<adv>$ ^*ӧстӱ/*ӧстӱ$, ^*бӱмӱк/*бӱмӱк$ ^ол<v><iv><past><p3><sg>/be<vblex><iv><past><p3><sg>/be able<vblex><iv><past><p3><sg>$, ^*җеннет/*җеннет$ ^*хапӱсӱн/*хапӱсӱн$ ^*ачтырды/*ачтырды$, ^*халхны/*халхны$ ^*йӱнахтан/*йӱнахтан$ ^*хутхарды/*хутхарды$.
 +
 
 +
=== Output of uum-eng (Full Translation) Mode ===
 +
"Start", - said - I did not start,
 +
the founder the star were born.
 +
Mary At #Bethlehem he did not find the peaceful place at.
 +
the village Joseph with he went at, the field the hay with he found the manger.
 +
The time came - Christ were born, *сарып the manger in he put.
 +
The day from the birth the light were born,
 +
from the light the star emerged,
 +
the shepherds  they frightened:
 +
"This how light this?"- they said.
 +
From The sky the angels descended, to the shepherds:
 +
"the relief brought, - The big God were born he said, -,
 +
*варың  the head" ,- bow he said.
 +
The angels, blessed they sang the holy song.
 +
The day from the birth the star came, *айдавҗусу he were the nightingale.
 +
The spark *йелир, the star fell, *чобаллара *мельэм he were, -
 +
At #Bethlehem Christ.
 +
were born Are *бешиге *эништилер, at the interior they saw Christ, *севиндилер to, the birth this *дӱнйайа *келишине.
 +
*Бӱйӱгӱ - *оҗа he saw,
 +
*ортанҗысы - *игит he saw,
 +
*эң *кичиги - *тахтын he found,
 +
to the head *пуллар *кийдирди.
 +
#What *ахыллы *халх they were,
 +
from, the sky from the star *бильдилер,
 +
the, myrrh brought the gold, the lamb they brought the gift.
 +
#Herod the king *ишитти, the army *алып he went to the path, hundred fourteen he killed thousand children, *Христозу  he did not find.
 +
The god *оғлуну *сахлады,
 +
*Ироднуң he tied the path.
 +
Christ went *Йегипитке, there *ӧстӱ, *бӱмӱк he were, *җеннет *хапӱсӱн *ачтырды, *халхны *йӱнахтан *хутхарды.
 +
 
 +
== Additions ==
 +
* '''Added 4 more (lexd) rules:'''
 +
** Conditional
 +
** Infinitive
 +
** Imperative
 +
** Added variations of existing rules (Dative, Locative, Ablative)
 +
* '''Added 4 more (twol) rules:'''
 +
** {з} gets realized as 'з' when preceding a vowel
 +
** {k} realizes as x after y or ы
 +
** н realizes as 'л' when preceding %{л%}
 +
** Added new category: 'silent'
 +
* '''Added 100+ more stems'''
 +
* '''Added 3 New Transfer Rules:'''
 +
** Corrected issue where proper noun names (i.e, Herod), did not analyze correctly.
 +
** Added article: 'the' to the transfer analysis.
 +
** Modified existential constructions вар, йох, and the "need" construction лазым.
 +
* '''Added apertium-separable rule to "merge" nouns ending with '-teen.''''
 +
 
 +
== Final Evaluation ==
 +
 
 +
=== (UUM) Transducer Evaluation ===
 +
==== Precision & Recall ====
 +
Precision: 55.67766%
 +
Recall: 73.07692%
 +
==== Coverage Over Large Corpus ====
 +
coverage: 172637 / 287386 (~0.60071471818390596619)
 +
 
 +
==== Total Words in Large Corpus ====
 +
287386
 +
 
 +
==== Total Stems in Transducer ====
 +
Lexicons: 49
 +
Lexicon entries: 424
 +
Patterns: 21
 +
Pattern entries: 45
 +
 
 +
Counts for individual lexicons:
 +
UNKNOWN: 52
 +
Punctuation: 22
 +
Adj-Root: 10
 +
Adv-Root: 12
 +
Conj-Root: 2
 +
Prep-Root: 2
 +
Part-Root: 14
 +
Det-Root: 1
 +
Intj-Root: 4
 +
Num-Root: 20
 +
Abbr-Root: 10
 +
Prn-Root: 17
 +
MetaTag: 2
 +
PropN-Root: 17
 +
Plurality: 3
 +
Possession: 6
 +
Cases: 9
 +
V-Root: 59
 +
TransitivityTag: 2
 +
Negation: 1
 +
VPres: 7
 +
VPresCond: 6
 +
VPast: 7
 +
VImp: 4
 +
VInf: 1
 +
VFutIndef: 7
 +
N-Root: 97
 +
Numeral: 10
 +
All anonymous lexicons: 2
 +
 
 +
=== (UUM ‣ ENG) Transfer Evaluation ===
 +
==== WER and PER Over Longer Corpus ====
 +
Test file: 'dev/tests/uum-eng.tests.longer.txt'
 +
Reference file '../ling073-uum-eng-corpus/eng.longer.txt'
 +
 
 +
Statistics about input files
 +
-------------------------------------------------------
 +
Number of words in reference: 278
 +
Number of words in test: 315
 +
Number of unknown words (marked with a star) in test: 7
 +
Percentage of unknown words: 2.22 %
 +
 
 +
Results when removing unknown-word marks (stars)
 +
-------------------------------------------------------
 +
Edit distance: 226
 +
Word error rate (WER): 81.29 %
 +
Number of position-independent correct words: 156
 +
Position-independent word error rate (PER): 57.19 %
 +
 
 +
Results when unknown-word marks (stars) are not removed
 +
-------------------------------------------------------
 +
Edit distance: 226
 +
Word Error Rate (WER): 81.29 %
 +
Number of position-independent correct words: 156
 +
Position-independent word error rate (PER): 57.19 %
 +
 
 +
Statistics about the translation of unknown words
 +
-------------------------------------------------------
 +
Number of unknown words which were free rides: 0
 +
Percentage of unknown words that were free rides: 0.00 %
 +
 
 +
==== Proportion of Stems Translated Correctly in Longer Corpus ====
  
−
Finally, for each of the 10 sentences you worked on, post on the wiki page the following:
+
Proportion of Stems Translated Correctly in Longer Corpus: 156 / 170 (~0.91764705882)
  
−
The original sentence,
+
==== Trimmed Coverage Over Longer & Large Corpora ====
  
−
The intended English translation,
+
Longer Corpora Coverage: 163 / 170 (~0.95882352941176470588)
  
−
The output of your xyz-abc-biltrans (lexical transfer) mode,
+
Large Corpora Coverage: 53127 / 186492 (~0.28487549063766810373)
  
−
The output of your xyz-abc (full translation) mode.
+
==== Number of Tokens in Longer and Large Corpora ====
 +
Longer Corpora Tokens: 170
  
 +
Large Corpora Tokens: 186492
  
 
[[Category:Sp23_TranslationPairs]][[Category:Urum]][[Category:English]]
 
[[Category:Sp23_TranslationPairs]][[Category:Urum]][[Category:English]]

Latest revision as of 21:53, 6 May 2023

Resources for Machine Translation Between Urum and English

Resources Developed for Urum

External resources

Transducer Links

Language Pair Repository

Transducer Coverage Over Example Text

Urum Transducer Coverage: 

coverage: 155 / 222 (~0.69819819819819819820)
remaining unknown forms: 67

Bilingual Transducer Coverage: 

coverage: 130 / 206 (~0.63106796116504854369)
remaining unknown forms: 76

(UUM ‣ ENG) Translation Evaluation

Urum Text

"Башла", - дэдим - башламадым,
башлавужу йылдыз доғду.
Панайана Вифлейемдэ тэна йер булмады кӧйд`е.
Йосиф илен т`етти чӧльд`е, пичен илен йасла булду.
Вахыт т`ельди - Христос доғду, сарып йасла чине хойду.
Кӱн доғуштан ышхын доғду,
ышихындан йылдыз кӧрӱндӱ,
чобанлар да хорхундулар:
"Бу нас йарых бу?"- дэдилер.
Кӧктэн мелайиклер энди, чобанлара гӧңӱль верди:
"Бӱйӱк аллах доғду, - дэди, -, 
варың да баш уруң" ,- дэди.
Мелайиклер хонуштулар, азиз тӱркӱ йырладылар.
Кӱн доғуштан йылдыз кельди, айдавҗусу бӱльбӱль эди.
Учхун йелир, йундуз хонду, чобаллара мельэм олду, -
Вифлейемдэ Христос доғду. 
Ол бешиге эништилер, ичиндэ Христос кӧрдӱлер, севиндилер доғушуна, бу дӱнйайа келишине.
Бӱйӱгӱ - оҗа кӧрдӱ,
ортанҗысы - игит кӧрдӱ,
эң кичиги - тахтын булду,
башына пуллар кийдирди.
Не ахыллы халх эдилер,
кӧктэн, йылдыздан бильдилер,
мирон, алтын т`етирдилер, хозу бағыш т`етирдилер.
Ирод падышах ишитти, аст`ер алып йола чыхты, йӱз он дӧрт бин бала хырды, Христозу да буламады.
Аллах оғлуну сахлады,
Ироднуң йолун бағлады.
Христос т`етти Йегипитке, онда ӧстӱ, бӱмӱк олду, җеннет хапӱсӱн ачтырды, халхны йӱнахтан хутхарды.

English Translation

"Start!" I said but did not start,
the dawn closed.
The Holy Mary in Bethlehem did not find a peaceful place in the village.
she went with Joseph to the field, found a manger with hay.
The time has come - Christ is born, wrapped and laid in a manger.
In the east, your light shone,
a star appeared in the light,
the shepherds were frightened:
"What is this light?" they said.
Angels descended from heavens, calmed the shepherds:
"A great God is born," they said,
"go and worship," they said.
Angels proclaimed, sang holy songs.
A star came from the east, accompanied by a nightingale.
The light approached, the star descended, the shepherds were comforted -
Christ was born in Bethlehem.
They descended to that crib, saw Christ in it, rejoiced at his birth, his coming into this world.
The eldest saw a husband,
the middle saw a knight,
the youngest found his throne,
they put glitter on his head.
What wise people they were -
they learned from the star in the sky,
they brought myrrh, gold, and a lamb as a gift.
King Herod searched, set out on the road with a star, killed fourteen thousand children, but couldn't find Christ.
Allah protected His son,
King Herod heard.
Christ went to Egypt, There He grew up and became great, opened the doors of paradise, saved people from sin.

Output of uum-eng-biltrans (Lexical Transfer) Mode

"^Башла<v><tv><imp><p1><sg>/Start<vblex><tv><imp><p1><sg>$", - ^дэ<v><iv><past><p1><sg>/say<vblex><iv><past><p1><sg>$ - ^башла<v><tv><neg><past><p1><sg>/start<vblex><tv><neg><past><p1><sg>$,
^башлавужу<n><nom>/founder<n><nom>$ ^йылдыз<n><nom>/star<n><nom>$ ^доғ<v><iv><past><p3><sg>/be# born<vblex><past><p3><sg>$.
^Панайана<np><ant><f><sg><nom>/Mary<np><ant><f><sg><nom>$ ^Вифлейем<np><top><sg><loc>/Bethlehem<np><top><sg><loc>$ ^тэна<adj>/peaceful<adj>$ ^йер<n><nom>/place<n><nom>$ ^бул<v><tv><neg><past><p3><sg>/find<vblex><tv><neg><past><p3><sg>$ ^кӧй<n><loc>/village<n><loc>$.
^Йосиф<np><ant><m><sg><nom>/Joseph<np><ant><m><sg><nom>$ ^илен<pr>/with<pr>$ ^тʼет<v><iv><past><p3><sg>/go<vblex><iv><past><p3><sg>$ ^чӧль<n><loc>/field<n><loc>$, ^пичен<n><nom>/hay<n><nom>$ ^илен<pr>/with<pr>$ ^йасла<n><nom>/manger<n><nom>$ ^бул<v><tv><past><p3><sg>/find<vblex><tv><past><p3><sg>$.
^Вахыт<n><nom>/Time<n><nom>$ ^тʼель<v><iv><past><p3><sg>/come<vblex><iv><past><p3><sg>$ - ^Христос<np><ant><m><sg><nom>/Christ<np><ant><m><sg><nom>$ ^доғ<v><iv><past><p3><sg>/be# born<vblex><past><p3><sg>$, ^*сарып/*сарып$ ^йасла<n><nom>/manger<n><nom>$ ^чине<pr>/in<pr>$ ^хой<v><tv><past><p3><sg>/put<vblex><tv><past><p3><sg>$.
^Кӱн<n><nom>/Sun<n><nom>/Day<n><nom>$ ^доғуш<n><abl>/birth<n><abl>$ ^ышхын<n><nom>/light<n><nom>$ ^доғ<v><iv><past><p3><sg>/be# born<vblex><past><p3><sg>$,
^ышхын<n><abl>/light<n><abl>$ ^йылдыз<n><nom>/star<n><nom>$ ^кӧрӱн<v><iv><past><p3><sg>/emerge<vblex><iv><past><p3><sg>$,
^чобан<n><pl><nom>/shepherd<n><pl><nom>$ ^да<part>/$ ^хорхун<v><iv><past><p3><pl>/frighten<vblex><past><p3><pl>$:
"^Бу<det>/This<det><dem><sg>$ ^нас<adv>/how<adv><itg>$ ^йарых<adj>/light<adj><sint>$ ^бу<det>/this<det><dem><sg>$?"- ^дэ<v><iv><past><p3><pl>/say<vblex><iv><past><p3><pl>$.
^Кӧк<n><abl>/Sky<n><abl>$ ^мелайик<n><pl><nom>/angel<n><pl><nom>$ ^эн<v><iv><past><p3><sg>/descend<vblex><iv><past><p3><sg>$, ^чобан<n><pl><dat>/shepherd<n><pl><dat>$ ^гӧңӱль<n><nom>/relief<n><nom>$ ^вер<v><tv><past><p3><sg>/bring<vblex><tv><past><p3><sg>$:
"^Бӱйӱк<adj>/Big<adj><sint>$ ^аллах<n><nom>/God<n><nom>$ ^доғ<v><iv><past><p3><sg>/be# born<vblex><past><p3><sg>$, - ^дэ<v><iv><past><p3><sg>/say<vblex><iv><past><p3><sg>$, -, 
^*варың/*варың$ ^да<part>/$ ^баш<n><nom>/head<n><nom>$ ^уруң<v><iv><imp><p1><sg>/bow<vblex><iv><imp><p1><sg>$" ,- ^дэ<v><iv><past><p3><sg>/say<vblex><iv><past><p3><sg>$.
^Мелайик<n><pl><nom>/Angel<n><pl><nom>$ ^хонуш<v><iv><past><p3><pl>/bless<vblex><iv><past><p3><pl>$, ^азиз<adj>/holy<adj><sint>$ ^тӱрк<n><acc>/song<n><acc>$ ^йырла<v><iv><past><p3><pl>/sing<vblex><iv><past><p3><pl>/йырла<vblex><iv><past><p3><pl>$.
^Кӱн<n><nom>/Sun<n><nom>/Day<n><nom>$ ^доғуш<n><abl>/birth<n><abl>$ ^йылдыз<n><nom>/star<n><nom>$ ^кель<v><iv><past><p3><sg>/come<vblex><iv><past><p3><sg>$, ^*айдавҗусу/*айдавҗусу$ ^бӱльбӱль<n><nom>/nightingale<n><nom>$ ^э<v><tv><past><p3><sg>/be<vblex><tv><past><p3><sg>$.
^Учхун<n><nom>/Spark<n><nom>$ ^*йелир/*йелир$, ^йундуз<n><nom>/star<n><nom>$ ^хон<v><iv><past><p3><sg>/fall<vblex><iv><past><p3><sg>$, ^*чобаллара/*чобаллара$ ^*мельэм/*мельэм$ ^ол<v><iv><past><p3><sg>/be<vblex><iv><past><p3><sg>/be able<vblex><iv><past><p3><sg>$, -
^Вифлейем<np><top><sg><loc>/Bethlehem<np><top><sg><loc>$ ^Христос<np><ant><m><sg><nom>/Christ<np><ant><m><sg><nom>$ ^доғ<v><iv><past><p3><sg>/be# born<vblex><past><p3><sg>$. 
^Ол<v><iv><imp><p1><sg>/Be<vblex><iv><imp><p1><sg>/Be able<vblex><iv><imp><p1><sg>$ ^*бешиге/*бешиге$ ^*эништилер/*эништилер$, ^ич<n><px3sg><loc>/stomach<n><px3sg><loc>/interior<n><px3sg><loc>$ ^Христос<np><ant><m><sg><nom>/Christ<np><ant><m><sg><nom>$ ^кӧр<v><tv><past><p3><pl>/see<vblex><tv><past><p3><pl>$, ^*севиндилер/*севиндилер$ ^доғуш<n><px3sg><dat>/birth<n><px3sg><dat>$, ^бу<det>/this<det><dem><sg>$ ^*дӱнйайа/*дӱнйайа$ ^*келишине/*келишине$.
^*Бӱйӱгӱ/*Бӱйӱгӱ$ - ^*оҗа/*оҗа$ ^кӧр<v><tv><past><p3><sg>/see<vblex><tv><past><p3><sg>$,
^*ортанҗысы/*ортанҗысы$ - ^*игит/*игит$ ^кӧр<v><tv><past><p3><sg>/see<vblex><tv><past><p3><sg>$,
^*эң/*эң$ ^*кичиги/*кичиги$ - ^*тахтын/*тахтын$ ^бул<v><tv><past><p3><sg>/find<vblex><tv><past><p3><sg>$,
^баш<n><px3sg><dat>/head<n><px3sg><dat>$ ^*пуллар/*пуллар$ ^*кийдирди/*кийдирди$.
^Не<prn><det><dem>/What<prn><det><dem>$ ^*ахыллы/*ахыллы$ ^*халх/*халх$ ^э<v><tv><past><p3><pl>/be<vblex><tv><past><p3><pl>$,
^кӧк<n><abl>/sky<n><abl>$, ^йылдыз<n><abl>/star<n><abl>$ ^*бильдилер/*бильдилер$,
^мирон<n><nom>/myrrh<n><nom>$, ^алтын<n><nom>/gold<n><nom>$ ^тʼетир<v><tv><past><p3><pl>/bring<vblex><tv><past><p3><pl>$, ^хоз<n><acc>/lamb<n><acc>$ ^бағыш<n><nom>/gift<n><nom>$ ^тʼетир<v><tv><past><p3><pl>/bring<vblex><tv><past><p3><pl>$.
^Ирод<np><ant><m><sg><nom>/Herod<np><ant><m><sg><nom>$ ^падышах<n><nom>/king<n><nom>$ ^*ишитти/*ишитти$, ^астʼер<n><nom>/army<n><nom>$ ^*алып/*алып$ ^йол<n><dat>/path<n><dat>$ ^чых<v><iv><past><p3><sg>/go<vblex><iv><past><p3><sg>/come<vblex><iv><past><p3><sg>$, ^йӱз<num>/hundred<num><pl>$ ^он дӧрт<num>/fourteen<num><pl>$ ^бин<num>/thousand<num><pl>$ ^бала<n><nom>/child<n><nom>$ ^хыр<v><tv><past><p3><sg>/kill<vblex><tv><past><p3><sg>$, ^*Христозу/*Христозу$ ^да<part>/$ ^була<v><tv><neg><past><p3><sg>/find<vblex><tv><neg><past><p3><sg>$.
^Аллах<n><nom>/God<n><nom>$ ^*оғлуну/*оғлуну$ ^*сахлады/*сахлады$,
^*Ироднуң/*Ироднуң$ ^йол<n><px3sg><nom>/path<n><px3sg><nom>$ ^бағла<v><tv><past><p3><sg>/tie<vblex><tv><past><p3><sg>$.
^Христос<np><ant><m><sg><nom>/Christ<np><ant><m><sg><nom>$ ^тʼет<v><iv><past><p3><sg>/go<vblex><iv><past><p3><sg>$ ^*Йегипитке/*Йегипитке$, ^онда<adv>/there<adv>$ ^*ӧстӱ/*ӧстӱ$, ^*бӱмӱк/*бӱмӱк$ ^ол<v><iv><past><p3><sg>/be<vblex><iv><past><p3><sg>/be able<vblex><iv><past><p3><sg>$, ^*җеннет/*җеннет$ ^*хапӱсӱн/*хапӱсӱн$ ^*ачтырды/*ачтырды$, ^*халхны/*халхны$ ^*йӱнахтан/*йӱнахтан$ ^*хутхарды/*хутхарды$.

Output of uum-eng (Full Translation) Mode

"Start", - said - I did not start,
the founder the star were born.
Mary At #Bethlehem he did not find the peaceful place at.
the village Joseph with he went at, the field the hay with he found the manger.
The time came - Christ were born, *сарып the manger in he put.
The day from the birth the light were born,
from the light the star emerged,
the shepherds  they frightened:
"This how light this?"- they said.
From The sky the angels descended, to the shepherds:
"the relief brought, - The big God were born he said, -, 
*варың  the head" ,- bow he said.
The angels, blessed they sang the holy song.
The day from the birth the star came, *айдавҗусу he were the nightingale.
The spark *йелир, the star fell, *чобаллара *мельэм he were, -
At #Bethlehem Christ. 
were born Are *бешиге *эништилер, at the interior they saw Christ, *севиндилер to, the birth this *дӱнйайа *келишине.
*Бӱйӱгӱ - *оҗа he saw,
*ортанҗысы - *игит he saw,
*эң *кичиги - *тахтын he found,
to the head *пуллар *кийдирди.
#What *ахыллы *халх they were,
from, the sky from the star *бильдилер,
the, myrrh brought the gold, the lamb they brought the gift.
#Herod the king *ишитти, the army *алып he went to the path, hundred fourteen he killed thousand children, *Христозу  he did not find.
The god *оғлуну *сахлады,
*Ироднуң he tied the path.
Christ went *Йегипитке, there *ӧстӱ, *бӱмӱк he were, *җеннет *хапӱсӱн *ачтырды, *халхны *йӱнахтан *хутхарды.

Additions

  • Added 4 more (lexd) rules:
    • Conditional
    • Infinitive
    • Imperative
    • Added variations of existing rules (Dative, Locative, Ablative)
  • Added 4 more (twol) rules:
    • {з} gets realized as 'з' when preceding a vowel
    • {k} realizes as x after y or ы
    • н realizes as 'л' when preceding %{л%}
    • Added new category: 'silent'
  • Added 100+ more stems
  • Added 3 New Transfer Rules:
    • Corrected issue where proper noun names (i.e, Herod), did not analyze correctly.
    • Added article: 'the' to the transfer analysis.
    • Modified existential constructions вар, йох, and the "need" construction лазым.
  • Added apertium-separable rule to "merge" nouns ending with '-teen.'

Final Evaluation

(UUM) Transducer Evaluation

Precision & Recall

Precision: 55.67766%
Recall: 73.07692%

Coverage Over Large Corpus

coverage: 172637 / 287386 (~0.60071471818390596619)

Total Words in Large Corpus

287386

Total Stems in Transducer

Lexicons: 49
Lexicon entries: 424
Patterns: 21
Pattern entries: 45

Counts for individual lexicons:
UNKNOWN: 52
Punctuation: 22
Adj-Root: 10
Adv-Root: 12
Conj-Root: 2
Prep-Root: 2
Part-Root: 14
Det-Root: 1
Intj-Root: 4
Num-Root: 20
Abbr-Root: 10
Prn-Root: 17
MetaTag: 2
PropN-Root: 17
Plurality: 3
Possession: 6
Cases: 9
V-Root: 59
TransitivityTag: 2
Negation: 1
VPres: 7
VPresCond: 6
VPast: 7
VImp: 4
VInf: 1
VFutIndef: 7
N-Root: 97
Numeral: 10
All anonymous lexicons: 2

(UUM ‣ ENG) Transfer Evaluation

WER and PER Over Longer Corpus

Test file: 'dev/tests/uum-eng.tests.longer.txt' Reference file '../ling073-uum-eng-corpus/eng.longer.txt'

Statistics about input files

Number of words in reference: 278 Number of words in test: 315 Number of unknown words (marked with a star) in test: 7 Percentage of unknown words: 2.22 %

Results when removing unknown-word marks (stars)

Edit distance: 226 Word error rate (WER): 81.29 % Number of position-independent correct words: 156 Position-independent word error rate (PER): 57.19 %

Results when unknown-word marks (stars) are not removed

Edit distance: 226 Word Error Rate (WER): 81.29 % Number of position-independent correct words: 156 Position-independent word error rate (PER): 57.19 %

Statistics about the translation of unknown words

Number of unknown words which were free rides: 0 Percentage of unknown words that were free rides: 0.00 %

Proportion of Stems Translated Correctly in Longer Corpus

Proportion of Stems Translated Correctly in Longer Corpus: 156 / 170 (~0.91764705882)

Trimmed Coverage Over Longer & Large Corpora

Longer Corpora Coverage: 163 / 170 (~0.95882352941176470588)

Large Corpora Coverage: 53127 / 186492 (~0.28487549063766810373)

Number of Tokens in Longer and Large Corpora

Longer Corpora Tokens: 170

Large Corpora Tokens: 186492
Retrieved from "http://wikis.swarthmore.edu/ling073/index.php?title=Urum_and_English&oldid=22294"