Korean Probabilistic Syntactic Model using Head Co-occurrence

摘要: Since a natural language has inherently structural ambiguities, one of the difficulties parsing is resolving ambiguities. Recently, probabilistic approach to tackle this disambiguation problem received considerable attention because it some attractions such as automatic learning, wide-coverage, and robustness. In paper, we focus on Korean model using head co-occurrence. We are apt meet data sparseness when we`re co-occurrence lexical. Therefore, how handle more important than others. To lighten problem, have used restricted simplified phrase-structure grammar back-off smoothing. The proposed showed that accuracy about 84%.