Magahi/Disambiguation

From LING073
Revision as of 22:09, 4 April 2021 by Pwheele1 (talk | contribs) (Repository)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search

Repository

Repository

Initial Evaluation of Ambiguity

./disambiguation-test.sh

basic corpus (mag.corpus.basic.txt):
Ambiguity before disambiguation: ~1.11354466858789625360
Ambiguity after disambiguation: ~1.04495677233429394813
full corpus (mag.corpus.full.txt):
Ambiguity before disambiguation: ~1.04468066337332392378
Ambiguity after disambiguation: ~1.03625617501764290755

Corpus

Forms in corpus with more than one analysis

cat corpus.txt | lt-proc /path/to/mag.automorf.bin | sed 's/$\W*\^/$\n^/g' | grep '\/.*\/'

Transducer

Number of all forms: 4293

hfst-expand mag.automorf.hfst | wc -l

Number of unique forms: 2985

hfst-expand mag.automorf.hfst | cut -f1 -d':' | sort -u | wc -l

Ambiguity of 1.43 (4293/2985).

Multiple Analyses:

hfst-expand mag.automorf.hfst | cut -f1 -d':' | sort > /tmp/totalforms
hfst-expand mag.automorf.hfst | cut -f1 -d':' | sort -u > /tmp/uniqforms
diff /tmp/totalforms /tmp/uniqforms

Sentences

  • ^laṛikā/laṛikā<n>/laṛikā<n><obl>/laṛikā<n><obl>$ ^i/<prn><dem>/<prn><dem><att>/<prn><pers><prx><p3><sg>$ ^likho/likh<v><pres><s_p3>/likh<v><pres><o_p2><hi>$^./.<sent>$
  • ^ham/*ham$ ^laṛikā/laṛikā<n>/laṛikā<n><obl>/laṛikā<n><obl>$ ^ke/ke<post>$ ^dekhi/dekh<v><pres><s_p1>$^./.<sent>$

Final Evaluation of Ambiguity

basic corpus (mag.corpus.basic.txt):
Ambiguity before disambiguation: ~1.06512968299711815562
Ambiguity after disambiguation: ~1.03400576368876080692

full corpus (mag.corpus.full.txt):
Ambiguity before disambiguation: ~1.04150494001411432604
Ambiguity after disambiguation: ~1.02999294283697953423