Difference between revisions of "Mirandese/Transducer"

Latest revision as of 21:30, 27 April 2023

Analyser Evaluation

• 165 stems (lexicon entries)
  • NounRoot: 11
  • RegNounInfl: 4
  • AdjRoot: 3
  • RegAdjInfl: 2
  • AdvRoot: 4
  • RegAdvInfl: 1
  • PrnRoot: 25
  • V-Stem: 3
  • Infin: 3
  • V-SimplePresentIndicative: 18
  • V-SimplePastIndicative: 18
  • V-SimpleFutureIndicative: 18
  • V-SimpleFutureSubjunctive: 18
  • Preposition: 1
  • Punctuation: 22
  • Numeral: 10
  • All anonymous lexicons: 4

TOP UNKNOWN WORDS:

    88 ^de/*de$
    48 ^i/*i$
    21 ^que/*que$
    16 ^ls/*ls$
    11 ^ye/*ye$
    11 ^a/*a$
    10 ^por/*por$
     9 ^ne/*ne$
     9 ^cumo/*cumo$
     9 ^an/*an$
     8 ^ua/*ua$
     8 ^se/*se$
     8 ^para/*para$
     8 ^na/*na$
     7 ^lhéngua/*lhéngua$
     6 ^un/*un$
     6 ^cun/*cun$
     5 ^ó/*ó$
     5 ^tecnologie/*tecnologie$
     5 ^ser/*ser$

coverage: 303 / 1279 (~0.23690383111806098514) remaining unknown forms: 976 ambiguity: 1321 / 1279 (~1.03283815480844409695)

Notes

By adding "de<pr> ↔ de" to the transducer, coverage went from 23.7% to 30.6%»

Generator Evaluation

Initial evaluation of morphological generation

Morph analysis tests:

Corpus 1 of 5: Nouns-morph

 11/11 (100.0%) tests pass (10/11 (90.91%) match gold)

Corpus 2 of 5: Adjectives-morph

 3/3 (100.0%) tests pass (0/3 (0.0%) match gold)

Corpus 3 of 5: Verbs-morph

 68/68 (100.0%) tests pass (30/68 (44.12%) match gold)

Corpus 4 of 5: Pronouns-morph

 22/22 (100.0%) tests pass (7/22 (31.82%) match gold)

Corpus 5 of 5: Adverbs-morph

 4/4 (100.0%) tests pass (4/4 (100.0%) match gold)

Current coverage number: 30.6%

Morph generator tests:

Corpus 1 of 5: Nouns-gen

 11/11 (100.0%) tests pass (5/11 (45.45%) match gold)

Corpus 2 of 5: Adjectives-gen

 3/3 (100.0%) tests pass (0/3 (0.0%) match gold)

Corpus 3 of 5: Verbs-gen

 72/72 (100.0%) tests pass (52/72 (72.22%) match gold)

Corpus 4 of 5: Pronouns-gen

 24/24 (100.0%) tests pass (13/24 (54.17%) match gold)

Corpus 5 of 5: Adverbs-gen

 3/3 (100.0%) tests pass (3/3 (100.0%) match gold)

Final evaluation of morphological generation

Corpus 1 of 5: Nouns-gen

 11/11 (100.0%) tests pass (11/11 (100.0%) match gold)

Corpus 2 of 5: Adjectives-gen

 3/3 (100.0%) tests pass (1/3 (33.33%) match gold)

Corpus 3 of 5: Verbs-gen

 72/72 (100.0%) tests pass (52/72 (72.22%) match gold)

Corpus 4 of 5: Pronouns-gen

 24/24 (100.0%) tests pass (13/24 (54.17%) match gold)

Corpus 5 of 5: Adverbs-gen

 3/3 (100.0%) tests pass (3/3 (100.0%) match gold)

• Four twol rules added.

• 30.6% coverage.