Dependency Analysis of Clauses Using Parse Tree Kernels

Abstract

Identification of dependency relation among clauses is one of the most critical parts in parsing Korean sentences because it generates severe ambiguities. The resolution of the ambiguities involves both syntactic and semantic information. This paper proposes a method to determine the dependency relation among Korean clauses using parse tree kernels. The parse tree used in this paper provides the method with the syntactic information, and the endings (Eomi) do with the semantic information. In addition, the parse tree kernel for handling parse trees has benefits that it minimizes the information loss occurred during transforming a parse tree into a feature vector, and can obtain, as a result, very accurate similarity between parse trees. The experimental results on a standard Korean data set show 89.12% of accuracy, which implies that the proposed method is plausible for the dependency analysis of clauses.KeywordsSupport Vector MachineNoun PhraseDependency RelationDependency AnalysisParse TreeThese keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.