Brazilian portuguese wordNet: a computational linguistic exercise of encoding bilingual relational lexicons

Abstract

This paper describes the methodology of encoding the Brazilian Portuguese WordNet (WN.Br) synsets and the automatic mapping of WN.Br´s conceptual relations of hyponymy, cohyponymy, meronymy, cause, and entailment relations from Princeton WordNet (WN.Pr). After contextualizing the project and outlining the current lexical database structure and its statistics, it is described the WN.Br editing tool to encode the synsets, its glosses and the equivalence EQ_RELATIONS between WN.Br and WN.Pr synsets, and to select sample sentences from corpora. The conclusion samples the automatic generation of WN.Br´s hyponymy and co-hyponymy conceptual relations from WN.Pr and outlines the ongoing work.

Artigo programático que apresenta ao graduando e ao pós-graduando o universo da construção do conhecimento sobre as línguas naturais com o auxílio de técnicas linguístico-computacionais.