Sardinian and English/Structural transfer

From LING073
Revision as of 19:39, 3 May 2023 by Rorlija1 (talk | contribs) (→‎.rtx rules)

Jump to: navigation, search

pre-evaluation

  • WER, PER, and other useful numbers:
 Statistics about input files
 -------------------------------------------------------
 Number of words in reference: 81
 Number of words in test: 80
 Number of unknown words (marked with a star) in test: 14
 Percentage of unknown words: 17.50 %
 Results when removing unknown-word marks (stars)
 -------------------------------------------------------
 Edit distance: 77
 Word error rate (WER): 95.06 %
 Number of position-independent correct words: 12
 Position-independent word error rate (PER): 85.19 %
  • coverage from monolingual transducer on srd.tests:
 Coverage: 97.78%
 Top unknown words in the corpus:
 1	   paìsu
 1	   Lettu
  • coverage from bilingual transducer on srd.tests:
 Number of tokenised words in the corpus: 91
 Coverage: 84.62%
 Top unknown words in the corpus:
 2	   bellu
 1	   paisu
 1	   paìsu
 1	   Lettu
 1	   furatu
 1	   dinari
 1	   punitos
 1	   rica
 1	   datu
 1	   milli
 1	   francos
 1	   betzu
 1	   iscimpru

.rtx rules

  • Item 1: Determiners before nouns in Sardinian are omitted in English translations that do not include determiners before nouns
    Ex: (srd) Sa Frantza est unu paisu bellu. → (eng) France is a beautiful country.
    Outputs for:
    tagger:
 ^Su<det><def><f><sg>$ ^Frantza<np><top><f><sg>$ ^èssere<vbser><pri><p3><sg>$ ^unu<det><ind><m><sg>$ ^paisu<n><m><sg>$ ^bellu<adj><m><sg>$^.<sent>$^.<sent>$
  • lex:
 1:SELECT:0.8:Su<det><def><f><sg>:2:The<det><def><f><sg>
 ^Su<det><def><f><sg>/The<det><def><f><sg>$ ^Frantza<np><top><f><sg>/France<np><top><f><sg>$ ^èssere<vbser><pri><p3><sg>/be<vbser><pri><p3><sg>$ ^unu<det><ind><m><sg>/a<det><ind><m><sg>$ ^paisu<n><m><sg>/country<n><sg>$ ^bellu<adj><m><sg>/beautiful<adj><m><sg>$^.<sent>/.<sent>$^.<sent>/.<sent>$
  • transfer:
 Applying rule 4 (line 58): ^Su<det><def><f><sg>/The<det><def><f><sg>$ ^Frantza<np><top><f><sg>/France<np><top><f><sg>$
 Applying rule 6 (line 61): ^èssere<vbser><pri><p3><sg>/be<vbser><pri><p3><sg>$
 Applying rule 1 (line 54): ^paisu<n><m><sg>/country<n><sg>$ ^bellu<adj><m><sg>/beautiful<adj><m><sg>$
 Applying rule 3 (line 57): ^unu<det><ind><m><sg>/a<det><ind><m><sg>$ ^default<NP><PD><ND>{^paisu<n><m><sg>/country<n><sg>$ ^bellu<adj><m><sg>/beautiful<adj><m><sg>$}$
 Applying rule 8 (line 64): ^be<vP><p3><sg>{^èssere<vbser><pri><p3><sg>/be<vbser><pri><p3><sg>$}$ ^default<DP><PD><ND>{^unu<det><ind><m><sg>/a<det><ind><m><sg>$ ^default<NP><PD><ND>{^paisu<n><m><sg>/country<n><sg>$ ^bellu<adj><m><sg>/beautiful<adj><m><sg>$}$}$
 Applying rule 9 (line 66): ^Default<DP><PD><ND>{^Su<det><def><f><sg>/The<det><def><f><sg>$ ^Frantza<np><top><f><sg>/France<np><top><f><sg>$}$ ^be<VP><p3><sg>{^be<vP><p3><sg>{^èssere<vbser><pri><p3><sg>/be<vbser><pri><p3><sg>$}$ ^default<DP><PD><ND>{^unu<det><ind><m><sg>/a<det><ind><m><sg>$ ^default<NP><PD><ND>{^paisu<n><m><sg>/country<n><sg>$ ^bellu<adj><m><sg>/beautiful<adj><m><sg>$}$}$}$
 Applying output rule 8 (line 66): Be<p3><sg> -> ^Default<DP><PD><ND>{^Su<det><def><f><sg>/The<det><def><f><sg>$ ^Frantza<np><top><f><sg>/France<np><top><f><sg>$}$ ^be<VP><p3><sg>{^be<vP><p3><sg>{^èssere<vbser><pri><p3><sg>/be<vbser><pri><p3><sg>$}$ ^default<DP><PD><ND>{^unu<det><ind><m><sg>/a<det><ind><m><sg>$ ^default<NP><PD><ND>{^paisu<n><m><sg>/country<n><sg>$ ^bellu<adj><m><sg>/beautiful<adj><m><sg>$}$}$}$
 Applying output rule 3 (line 58): Default<DP><p3><sg> -> ^Su<det><def><f><sg>/The<det><def><f><sg>$ ^Frantza<np><top><f><sg>/France<np><top><f><sg>$
 No rule specified: ^France<np><top><sg>$
 Applying output rule 7 (line 64): be<VP><p3><sg> -> ^be<vP><p3><sg>{^èssere<vbser><pri><p3><sg>/be<vbser><pri><p3><sg>$}$ ^default<DP><PD><ND>{^unu<det><ind><m><sg>/a<det><ind><m><sg>$ ^default<NP><PD><ND>{^paisu<n><m><sg>/country<n><sg>$ ^bellu<adj><m><sg>/beautiful<adj><m><sg>$}$}$
 Applying output rule 5 (line 61): be<vP><p3><sg> -> ^èssere<vbser><pri><p3><sg>/be<vbser><pri><p3><sg>$
 No rule specified: ^be<vbser><pres><p3><sg>$
 Applying output rule 2 (line 57): default<DP><p3><sg> -> ^unu<det><ind><m><sg>/a<det><ind><m><sg>$ ^default<NP><PD><ND>{^paisu<n><m><sg>/country<n><sg>$ ^bellu<adj><m><sg>/beautiful<adj><m><sg>$}$
 No rule specified: ^a<det><ind><sg>$
 Applying output rule 0 (line 54): default<NP><p3><sg> -> ^paisu<n><m><sg>/country<n><sg>$ ^bellu<adj><m><sg>/beautiful<adj><m><sg>$
 No rule specified: ^beautiful<adj>$
 No rule specified: ^country<n><sg>$
 No rule specified: ^.<sent>/.<sent>$
 No rule specified: ^.<sent>/.<sent>$
 ^France<np><top><sg>$ ^be<vbser><pres><p3><sg>$ ^a<det><ind><sg>$ ^beautiful<adj>$ ^country<n><sg>$^.<sent>$^.<sent>$
  • abc-xyz:
 France is a beautiful country.
  • Item 2: Noun-adjective order in noun phraseology: adjectives now come before the noun in translations when necessary:
    Ex: (srd) Su sardu est una limba antica. → (eng) Sardinian is an ancient language.
    Outputs for:
    tagger:
 ^Su<det><def><m><sg>$ ^sardu<n><m><sg>$ ^èssere<vbser><pri><p3><sg>$ ^unu<det><ind><f><sg>$ ^limba<n><f><sg>$ ^anticu<adj><f><sg>$^.<sent>$^.<sent>$
  • lex:
 1:SELECT:0.8:Su<det><def><m><sg>:2:The<det><def><m><sg>
 ^Su<det><def><m><sg>/The<det><def><m><sg>$ ^sardu<n><m><sg>/Sardinian<n><sg>$ ^èssere<vbser><pri><p3><sg>/be<vbser><pri><p3><sg>$ ^unu<det><ind><f><sg>/a<det><ind><f><sg>$ ^limba<n><f><sg>/language<n><sg>$ ^anticu<adj><f><sg>/ancient<adj><f><sg>$^.<sent>/.<sent>$^.<sent>/.<sent>$
  • transfer:
 No rule specified: ^Su<det><def><m><sg>/The<det><def><m><sg>$
 No rule specified: ^sardu<n><m><sg>/Sardinian<n><sg>$
 Applying rule 6 (line 61): ^èssere<vbser><pri><p3><sg>/be<vbser><pri><p3><sg>$
 Applying rule 1 (line 54): ^limba<n><f><sg>/language<n><sg>$ ^anticu<adj><f><sg>/ancient<adj><f><sg>$
 Applying rule 3 (line 57): ^unu<det><ind><f><sg>/a<det><ind><f><sg>$ ^default<NP><PD><ND>{^limba<n><f><sg>/language<n><sg>$ ^anticu<adj><f><sg>/ancient<adj><f><sg>$}$
 Applying rule 8 (line 64): ^be<vP><p3><sg>{^èssere<vbser><pri><p3><sg>/be<vbser><pri><p3><sg>$}$ ^default<DP><PD><ND>{^unu<det><ind><f><sg>/a<det><ind><f><sg>$ ^default<NP><PD><ND>{^limba<n><f><sg>/language<n><sg>$ ^anticu<adj><f><sg>/ancient<adj><f><sg>$}$}$
 Applying rule 10 (line 67): ^be<VP><p3><sg>{^be<vP><p3><sg>{^èssere<vbser><pri><p3><sg>/be<vbser><pri><p3><sg>$}$ ^default<DP><PD><ND>{^unu<det><ind><f><sg>/a<det><ind><f><sg>$ ^default<NP><PD><ND>{^limba<n><f><sg>/language<n><sg>$ ^anticu<adj><f><sg>/ancient<adj><f><sg>$}$}$}$
 Applying output rule 9 (line 67): be<S><p3><sg> -> ^be<VP><p3><sg>{^be<vP><p3><sg>{^èssere<vbser><pri><p3><sg>/be<vbser><pri><p3><sg>$}$ ^default<DP><PD><ND>{^unu<det><ind><f><sg>/a<det><ind><f><sg>$ ^default<NP><PD><ND>{^limba<n><f><sg>/language<n><sg>$ ^anticu<adj><f><sg>/ancient<adj><f><sg>$}$}$}$
 No rule specified: ^prpers<prn><subj><p3><nt><sg>$
 Applying output rule 7 (line 64): be<VP><p3><sg> -> ^be<vP><p3><sg>{^èssere<vbser><pri><p3><sg>/be<vbser><pri><p3><sg>$}$ ^default<DP><PD><ND>{^unu<det><ind><f><sg>/a<det><ind><f><sg>$ ^default<NP><PD><ND>{^limba<n><f><sg>/language<n><sg>$ ^anticu<adj><f><sg>/ancient<adj><f><sg>$}$}$
 Applying output rule 5 (line 61): be<vP><p3><sg> -> ^èssere<vbser><pri><p3><sg>/be<vbser><pri><p3><sg>$
 No rule specified: ^be<vbser><pres><p3><sg>$
 Applying output rule 2 (line 57): default<DP><p3><sg> -> ^unu<det><ind><f><sg>/a<det><ind><f><sg>$ ^default<NP><PD><ND>{^limba<n><f><sg>/language<n><sg>$ ^anticu<adj><f><sg>/ancient<adj><f><sg>$}$
 No rule specified: ^a<det><ind><sg>$
 Applying output rule 0 (line 54): default<NP><p3><sg> -> ^limba<n><f><sg>/language<n><sg>$ ^anticu<adj><f><sg>/ancient<adj><f><sg>$
 No rule specified: ^ancient<adj>$
 No rule specified: ^language<n><sg>$
 No rule specified: ^.<sent>/.<sent>$
 No rule specified: ^.<sent>/.<sent>$
 ^The<det><def><m><sg>$ ^Sardinian<n><sg>$ ^prpers<prn><subj><p3><nt><sg>$ ^be<vbser><pres><p3><sg>$ ^a<det><ind><sg>$ ^ancient<adj>$ ^language<n><sg>$^.<sent>$^.<sent>$
  • abc-xyz:
 #The Sardinian it is an ancient language.
  • Item 2: Posessive determiner/marker order in noun phraseology
    Ex: (srd) Libru meu est. → (eng) It is my book.
    Outputs for:
    tagger:^Libru<n><m><sg>$ ^meu<adj><pos><m><sg>$ ^est<n><m><sg>$^.<sent>$^.<sent>$
    lex
    transfer:
    abc-xyz:
  • Item 3: The English determiner "the" is immutable, while the Sardinian determiner "su" must agree in number and gender with the noun it modifies
    Ex: (srd) Una de sas berbekes est maláida. → (eng) One of the sheep is sick.
    Outputs for:
    tagger:
    lex
    transfer:
    abc-xyz:
  • Item 4: For most uses/conjugations of verbs in Sardinian, the verb stands on its own without a personal pronoun, while in English personal pronouns always precede any conjugated verbs. In other words, the tags of "appo" (I have) determine a new word "I" which precedes "have" in the English translation.
    Ex: (srd) Appo pèrditu duos de sos libros meos. → (eng) I have lost two of my books.
    Outputs for:
    tagger:
    lex
    transfer:
    abc-xyz:

post evaluation

  • WER, PER, and other useful numbers, part 2:
  • Coverage from monolingual transducer, part 2:
  • Coverage from bilingual transducer, part 2:,,