Evaluation of a Japanese CFG Derived from a Syntactically Annotated Corpus with Respect to Dependency Measures
Tomoya Noro, Chimato Koike, Taiichi Hashimoto, Takenobu Tokunaga, Hozumi Tanaka · Tokyo Tech Research Repository (Tokyo Institute of Technology) · 2005
Parsing is one of the important processes for natural language processing and, in general, a large-scale CFG is used to parse a wide variety of sentences. For many languages, a CFG is derived from a large-scale syntactically annotated corpus, and many parsing algorithms using CFGs have been proposed. However, we could not apply them to Japanese since a Japanese syntactically annotated corpus has not been available as of yet. In order to solve the problem, we have been building a large-scale Japanese syntactically annotated corpus. In this paper, we show the evaluation results of a CFG derived from our corpus and compare it with results of some Japanese dependency analyzers.