Semantic Annotation - VISL
Semantic Annotation - VISL
Semantic Annotation - VISL
Create successful ePaper yourself
Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.
A usage example: Using “deeply” annotated<br />
corpora for “live” lexicography: DeepDict<br />
www.gramtrans.com<br />
1) annotate corpora with dependency, syntactic function<br />
and semantic classes<br />
– Linguateca's Público and Folha corpora (CETEMPúblico,<br />
CETENFolha) – ca. 180+30 M words<br />
– Portuguese Wikipedia (Nov. 2005) – ca. 8 M words<br />
– Portuguese section of Europarl – ca. 27 M words<br />
2) extract dependency pairs of word + complements<br />
– N + @N< (vaca louca), @>A + ADJ (gravemente doente)<br />
– V + @ACC (ganhar terreno), @SUBJ ktp, V + PRP (pensar em)<br />
3) identify statistically significant relations (not n-<br />
grams but “dep-grams!”): log (p(AB)^2 / (p(A) * p (B)))