SYNGRAPH: A Flexible Matching Method based on Synonymous Expression Extraction from an Ordinary Dictionary and a Web Corpus.
Tomohide Shibata, Michitaka Odani, Jun Harashima, Takashi Oonishi, Sadao Kurohashi · 2008
This paper proposes a flexible matching method that can assimilate the expressive divergence. First, broad-coverage synonymous expressions are automatically extracted from an ordinary dictionary, and among them, those whose distributional similarity in a Web corpus is high are used for the flexible matching. Then, to overcome the combinatorial explosion problem in the combination of expressive divergence, an ID is assigned to each synonymous group, and SYNGRAPH data structure is introduced to pack the expressive divergence. We confirmed the effectiveness of our method on experiments of machine translation and information retrieval. 1