Valence Induction with a Head-Lexicalized PCFG

Abstract

This paper presents an experiment in learning valences (subcategorization frames) from a 50 million word text corpus, based on a lexicalized probabilistic context free grammar. Distributions are estimated using a modified EM algorithm. We evaluate the acquired lexicon both by comparison with a dictionary and by entropy measures. Results show that our model produces highly accurate frame distributions.

10 pages, 5 postscript figures

Cited by in corpus (7)

• Parameter Learning of Logic Programs for Symbolic-Statistical Modeling
• Can Subcategorisation Probabilities Help a Statistical Parser?
• Exploiting auxiliary distributions in stochastic unification-based grammars
• Valence extraction using EM selection and co-occurrence matrices
• Automatic Extraction of Subcategorization Frames for Czech
• Inside-Outside Estimation of a Lexicalized PCFG for German
• Using a Probabilistic Class-Based Lexicon for Lexical Ambiguity Resolution