User talk:Msalperen

From Apertium

[edit] Notes

• The verbs "Ovalamak" and "Kovalamak" creates some problems. This is about ambiguity which is expected to be solved by apertium tagger. Ova (champaign), ovalamak (to scrub) so ovalarım both means my champaigns and I scrub :D (likewise kovalarım both means "my buckets" and "I chase")

^ovalarım/ova<n><pl><pp1s><abs>/ovalamak<vblex><aff><pres><p1><sg><stat>$ ^kovalarım/kova<n><pl><pp1s><abs>/kovalamak<vblex><aff><pres><p1><sg><stat>$ apertium-tagger

^ova<n><pl><pp1s><abs>$ ^kova<n><pl><pp1s><abs>$ apertium-pretransfer

^ova<n><pl><pp1s><abs>$ ^kova<n><pl><pp1s><abs>$ apertium-transfer

^nom<NP><pl>{^n><n><pl><pp1s><abs>$}$ ^nom<NP><pl>{^vedrə<n><pl><pp1s><abs>$}$ apertium-interchunk

^nom<NP><pl>{^n><n><pl><pp1s><abs>$}$ ^nom<NP><pl>{^vedrə<n><pl><pp1s><abs>$}$ apertium-postchunk

^n><n><pl><pp1s><abs>$ ^vedrə<n><pl><pp1s><abs>$ lt-proc

1. n> vedrələrim

lt-proc

1. n> vedrələrim
2. n> vedrələrim

[edit] To Do List

• Inflectional morphology for verbs

• Inflectional morphology for nouns (cases)

• Derivational morphology for verbs

• Derivational morphology for nouns

[edit] Notes

Statistical Machine Translation on Paraphrased Corpora

• http://www.kecl.ntt.co.jp/icl/kpro/taro/papers/lrec02-normalized.pdf

Artificial Data

• http://www.cs.cmu.edu/afs/cs/project/jair/pub/volume18/thompson03a-html/node19.html

Turkish wordlist (also English and Finnish)

• http://www.cis.hut.fi/morphochallenge2005/datasets.shtml

• Paraphrasing:

http://www.coli.uni-saarland.de/projects/igk/annualMeeting/contents/2004/Projects/paraphrases/

• SMT with Prolog:

sourceforge.net/projects/psmt/

• SMT Quick Run:

http://ufal.mff.cuni.cz/~curin/SMT_QuickRun/