Difference between revisions of "Miskito and English/Structural transfer"
(→Plural Pronouns) |
(→Post-Evaluation) |
||
Line 78: | Line 78: | ||
=Post-Evaluation= | =Post-Evaluation= | ||
+ | |||
+ | Test file: 'miq-eng.tests.txt' | ||
+ | Reference file '../ling073-miq-eng-corpus/eng.tests.txt' | ||
+ | |||
+ | Statistics about input files | ||
+ | |||
+ | ------------------------------------------------------- | ||
+ | |||
+ | Number of words in reference: 41 | ||
+ | |||
+ | Number of words in test: 38 | ||
+ | |||
+ | Number of unknown words (marked with a star) in test: 3 | ||
+ | Percentage of unknown words: 7.89 % | ||
+ | |||
+ | Results when removing unknown-word marks (stars) | ||
+ | ------------------------------------------------------- | ||
+ | Edit distance: 35 | ||
+ | Word error rate (WER): 85.37 % | ||
+ | Number of position-independent correct words: 14 | ||
+ | Position-independent word error rate (PER): 65.85 % | ||
+ | |||
+ | Results when unknown-word marks (stars) are not removed | ||
+ | ------------------------------------------------------- | ||
+ | Edit distance: 35 | ||
+ | Word Error Rate (WER): 85.37 % | ||
+ | Number of position-independent correct words: 14 | ||
+ | Position-independent word error rate (PER): 65.85 % | ||
+ | |||
+ | 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 % |
Latest revision as of 14:09, 12 May 2021
Contents
Pre-Evaluation
Test file: 'miq-eng.tests.txt' Reference file '../ling073-miq-eng-corpus/eng.tests.txt'
Statistics about input files
Number of words in reference: 17
Number of words in test: 15
Number of unknown words (marked with a star) in test: 10
Percentage of unknown words: 66.67 %
Results when removing unknown-word marks (stars)
Edit distance: 16
Word error rate (WER): 94.12 %
Number of position-independent correct words: 1
Position-independent word error rate (PER): 94.12 %
Results when unknown-word marks (stars) are not removed
Edit distance: 16
Word Error Rate (WER): 94.12 %
Number of position-independent correct words: 1
Position-independent word error rate (PER): 94.12 %
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 %
Example Sentences
Plural Nouns
(miq) Yang bîp nani kaikri. | (eng) I saw the cows.
Tagger
^Yang<prn><p1>$ ^bîp<n>$ ^nani<det><pl>$ ^kaikaia<v><past><p1>$^.<sent>$^.<sent>$
Lex
^Yang<prn><p1>/Prpers<prn><CD><p1><mf><sg>$ ^bîp<n>/cow<n>$ ^nani<det><pl>/$ ^kaikaia<v><past><p1>/see<vblex><past><p1>$^.<sent>/.<sent>$^.<sent>/.<sent>$
Transfer
^Prpers<prn><obj><p1><mf><sg>$ ^cow<n><pl>$ ^see<vblex><past><p1>$^.<sent>$^.<sent>$
Biltrans
^Yang<prn><p1>/Prpers<prn><CD><p1><mf><sg>$ ^bîp<n>/cow<n>$ ^nani<det><pl>/$ ^kaikaia<v><past><p1>/see<vblex><past><p1>$^.<sent>/.<sent>$^.<sent>/.<sent>$
Miq-Eng
Me cows #see.
Plural Pronouns
(miq) Witin nani pain sa. | (eng) They are good.
Tagger
^Witin<prn><p3>$ ^nani<det><pl>$ ^pain<adj>$ ^kaia<vkaia><pres><p3>$^.<sent>$^.<sent>$
Lex
^Witin<prn><p3>/Prpers<prn><CD><p3><sg>$ ^nani<det><pl>/$ ^pain<adj>/good<adj>$ ^kaia<vkaia><pres><p3>/be<vbser><pres><p3>$^.<sent>/.<sent>$^.<sent>/.<sent>$
Transfer
^Prpers<prn><obj><p3><m><pl>$ ^good<adj>$ ^be<vbser><pres><p3>$^.<sent>$^.<sent>$
Biltrans
^Witin<prn><p3>/Prpers<prn><CD><p3><sg>$ ^nani<det><pl>/$ ^pain<adj>/good<adj>$ ^kaia<vkaia><pres><p3>/be<vbser><pres><p3>$^.<sent>/.<sent>$^.<sent>/.<sent>$
Miq-Eng
Them #good #be.
Post-Evaluation
Test file: 'miq-eng.tests.txt' Reference file '../ling073-miq-eng-corpus/eng.tests.txt'
Statistics about input files
Number of words in reference: 41
Number of words in test: 38
Number of unknown words (marked with a star) in test: 3 Percentage of unknown words: 7.89 %
Results when removing unknown-word marks (stars)
Edit distance: 35 Word error rate (WER): 85.37 % Number of position-independent correct words: 14 Position-independent word error rate (PER): 65.85 %
Results when unknown-word marks (stars) are not removed
Edit distance: 35 Word Error Rate (WER): 85.37 % Number of position-independent correct words: 14 Position-independent word error rate (PER): 65.85 %
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 %