A hybird method for automatically tagging Chinese sentences

Abstract

Automatically tagging Part-of-Speech is an important task in natural language processing. A better tagging result can increase the accuracy and performance of sentences parsing and semantics analysis. The major problem of the tagging is that the ambiguity in multiple-tag words caused the accuracy of the tagging decrease. Many approaches were proposed to solve the problem. However, some drawbacks still existed in these methods. In this paper, according to Chinese linguistic properties, we presented a hybrid method from two-way adjoining relation method and temporally clustering modification method for improving tagging. In preliminary experiments, it can really increase the accurate rate of the tagging for Chinese.