Magahi and English/Structural transfer

From LING073
Revision as of 19:37, 3 May 2021 by Pwheele1 (talk | contribs) (Verb Conjugation)

Jump to: navigation, search

Pre-Evaluation

Statistics about input files


Number of words in reference: 52 Number of words in test: 42 Number of unknown words (marked with a star) in test: 10 Percentage of unknown words: 23.81 %

Results when removing unknown-word marks (stars)


Edit distance: 52 Word error rate (WER): 100.00 % Number of position-independent correct words: 2 Position-independent word error rate (PER): 96.15 %

Examples

Postpositions

Tagger:

^pāoṃ<n>$ ^lāg<v><pres>$ ^ke<post>$^.<sent>$^.<sent>

Lex:

Transfer:


mag-eng:


Verb Conjugation

(mag) Ham sutli. → (eng) I slept.
Tagger:

^Prpers<prn><pers><p1><sg>$ ^*sutli$^.<sent>$

Lex:

^Prpers<prn><pers><p1><sg>/Prpers<prn><pers><p1><sg>$ ^sut<v><past><s_p1>/sleep<

vblex><past><s_p1>$^.<sent>/.<sent>$^.<sent>/.<sent>

Transfer:

No rule specified: ^Prpers<prn><pers><p1><sg>/Prpers<prn><pers><p1><sg>$
Applying rule 9 (line 64): ^sut<v><past><s_p1>/sleep<vblex><past><s_p1>$
Applying rule 8 (line 62): ^default<vP><s_p1>{^sut<v><past><s_p1>/sleep<vblex><past><s_p1>$}$
Applying output rule 7 (line 62): default<VP><s_p1> -> ^default<vP><s_p1>{^sut<v><past><s_p1>/sleep<vblex><past><s_p1>$}$
Applying output rule 8 (line 64): default<vP><s_p1> -> ^sut<v><past><s_p1>/sleep<vblex><past><s_p1>$
No rule specified: ^sleep<vblex><past>$
No rule specified: ^.<sent>/.<sent>$
No rule specified: ^.<sent>/.<sent>$
^Prpers<prn><pers><p1><sg>$ ^sleep<vblex><past>$^.<sent>$^.<sent>

mag-eng:

I slept

Verb Tenses

(mag) Sādhū dekhbo. → (eng) The saint will slee. Tagger:

^Sādhū<n>$ ^dekh<v><fut><s_p3>$^.<sent>$

Lex:

Transfer:


mag-eng:

The saint will see.

Post-Evaluation

Number of words in test: 42 Number of unknown words (marked with a star) in test: 10 Percentage of unknown words: 23.81 %

Results when removing unknown-word marks (stars)


Edit distance: 50 Word error rate (WER): 84.75 % Number of position-independent correct words: 9 Position-independent word error rate (PER): 84.75 %

Results when unknown-word marks (stars) are not removed


Edit distance: 59 Word Error Rate (WER): 100.00 % Number of position-independent correct words: 0 Position-independent word error rate (PER): 100.00 %