Difference between revisions of "Jeju and Ainu"
From LING073
(→Final evaluation) |
(→jje->ain) |
||
| Line 73: | Line 73: | ||
====Transducer==== | ====Transducer==== | ||
| − | *Precision and Recall against basic corpus: | + | *Precision and Recall against basic corpus: Precision 29.44785% / Recall 41.37931% |
*Coverage: | *Coverage: | ||
*Number of words in the large corpus | *Number of words in the large corpus | ||
Revision as of 03:17, 20 April 2017
Contents
Resources for machine translation between Jeju and Ainu
- Both translation directions are at: ling073-ain-jje
- (Deprecated): ling073-jje-ain
ain → jje evaluation
tests.txt evaluation
- WER: 100%
- PER: 100%
- Proportion of stems translated correctly: 11/11
sentences.txt evaluation
- WER: 100%
- PER: 100%
- Proportion of stems translated correctly: 73/296
jje → ain evaluation
tests.txt evaluation
- WER: 100%
- PER: 100%
- Proportion of stems translated correctly: 8/14
sentences.txt evaluation
- WER: 177.98% (not sure how these error rates can be above 100% but somehow we managed it)
- PER: 177.98%
- Proportion of stems translated correctly: 150/502
Lexical Selection
jje -> ain
- (jje)눈 -> (ain) ウパㇱ (snow)
- (jje)눈 -> (ain) シㇰ (eye)
- (jje)배 -> (ain) チㇷ゚ (ship)
- (jje)배 -> (ain) ト゚ィ (stomach)
ain -> jje
- (ain) サポ (older sister) -> 누나 (older sister speaker:male)
- (ain) サポ (older sister) -> 언니 (older sister, speaker:female)
- (ain) チュㇷ゚ (sun/moon/month) -> 달 (moon/month)
- (ain) チュㇷ゚ (sun/moon/month) -> 해 (sun)
Developed resources
Contrastive grammar
Structural transfer
Final evaluation
ain->jje
Transducer
- Precision and Recall against basic corpus:
- Coverage:
- Number of words in the large corpus
- Number of stems i the transducer
Machine Translation
- WER and PER over large corpus
- Proportion of stems translated correctly in large corpus
- Trimmed coverage over large corpus
- Number of stems in large corpus
jje->ain
Implemented
- 200 more stems to the bilingual dictioanry
- Expanded morphology
- Lexical selection rules
Transducer
- Precision and Recall against basic corpus: Precision 29.44785% / Recall 41.37931%
- Coverage:
- Number of words in the large corpus
- Number of stems i the transducer
Machine Translation
- WER and PER over large corpus
- Proportion of stems translated correctly in large corpus
- Trimmed coverage over large corpus
- Number of stems in large corpus
