Difference between revisions of "Waray/Transducer"
From LING073
(→Notes) |
(→Generator Evaluation) |
||
Line 56: | Line 56: | ||
==Generator Evaluation== | ==Generator Evaluation== | ||
+ | ===Initial Evaluation=== | ||
+ | ====Analysis==== | ||
+ | <code>morph-test -csi tests/war.yaml</code> | ||
+ | * Total passes: 55, Total fails: 24, Total: 79 | ||
− | + | ====Generation==== | |
+ | <code>morph-test -cl tests/war.yaml</code> | ||
+ | * Total passes: 55, Total fails: 19, Total: 74 |
Revision as of 16:09, 20 April 2021
Contents
Morphological Transducer
Analyser Evaluation
war.corpus Sun 18 Apr 2021 04:10:45 AM EDT 592/1259 ~0.47021445591739475774
- Coverage: 592/1259 (~0.47021445591739475774)
- Remaining unknown forms: 667
Lexical Information
- Lexicons: 15
- Lexicon entries: 112
- Patterns: 1
- Pattern entries: 8
Counts for individual lexicon
- NounRoot: 19
- Determiners: 11
- PluralDet: 2
- VerbPrefixes: 5
- VerbStems: 5
- Pronouns: 31
- Adverbs: 11
- Auxiliary: 1
- All anonymous lexicons: 5
Tests
- war.yaml: Total passes: 56, Total fails: 15, Total: 71
Current Unknown Words
TOP UNKNOWN WORDS:
94 ^nga/*nga$ 28 ^tuna/*tuna$ 15 ^nagkaada/*nagkaada$ 15 ^Dios/*Dios$ 12 ^hito/*hito$ 11 ^uyon/*uyon$ 11 ^linarang/*linarang$ 11 ^adlaw/*adlaw$ 10 ^may/*may$ 10 ^langit/*langit$ 10 ^i/*i$ 9 ^klase/*klase$ 9 ^hayop/*hayop$ 9 ^gab/*gab$ 8 ^katubigan/*katubigan$ 7 ^tagsa/*tagsa$ 7 ^ngatanan/*ngatanan$ 7 ^kalamrag/*kalamrag$ 7 ^basi/*basi$ 6 ^nalupad/*nalupad$
Notes
- The unknown words gab and i are likely to be from the same word gab-i (night)
- Tests for verbalized nouns not implemented yet
- Disambiguation still needed
Generator Evaluation
Initial Evaluation
Analysis
morph-test -csi tests/war.yaml
- Total passes: 55, Total fails: 24, Total: 79
Generation
morph-test -cl tests/war.yaml
- Total passes: 55, Total fails: 19, Total: 74