Capturing consistency between intra-clause and inter-clause relations in knowledge-rich dependency and case structure analysis

We present a method for dependency and case structure analysis that captures the consistency between intra-clause relations (i.e., case structures or predicate-argument structures) and inter-clause relations. We assess intra-clause relations on the basis of case frames and inter-clause relations on the basis of transition knowledge between case frames. Both knowledge bases are automatically acquired from a massive amount of parses of a Web corpus. The significance of this study is that the proposed method selects the best dependency and case structure that are consistent within each clause and between clauses. We confirm that this method contributes to the improvement of dependency parsing of Japanese.