Paper: Capturing Paradigmatic and Syntagmatic Lexical Relations: Towards Accurate Chinese Part-of-Speech Tagging

ACL ID P12-1026
Title Capturing Paradigmatic and Syntagmatic Lexical Relations: Towards Accurate Chinese Part-of-Speech Tagging
Venue Annual Meeting of the Association of Computational Linguistics
Session Main Conference
Year 2012
Authors

From the perspective of structural linguistics, we explore paradigmatic and syntagmatic lex- ical relations for Chinese POS tagging, an im- portant and challenging task for Chinese lan- guage processing. Paradigmatic lexical rela- tions are explicitly captured by word cluster- ing on large-scale unlabeled data and are used to design new features to enhance a discrim- inative tagger. Syntagmatic lexical relations are implicitly captured by constituent pars- ing and are utilized via system combination. Experiments on the Penn Chinese Treebank demonstrate the importance of both paradig- matic and syntagmatic relations. Our linguis- tically motivated approaches yield a relative error reduction of 18% in total over a state- of-the-art baseline.