作者: Drahomíra "johanka" Spoustová , Jan Hajič , Jan Votrubec , Pavel Krbec , Pavel Květoň
关键词:
摘要: Several hybrid disambiguation methods are described which combine the strength of hand-written rules and statistical taggers. Three different (HMM, Maximum-Entropy Averaged Perceptron) taggers used in a tagging experiment using Prague Dependency Tree-bank. The results systems better than any other method tried for Czech so far.