汉语分词和词性标注一体化分析的方法研究

Research on the Approach of Integratng Chinese Word Segmentation with Part-of-speech Tagging

提出了一种汉语文本切分和词性标注相融合的一体化分析的统计模型，并应用动态规划算法与Ａ＊解码算法相结合的二次搜索算法，实现了一个基于该模型的汉语词法分析器。初步的开放测试表明，该分析器的分词准确率和词性标注正确率分别可达９８．６７％和９５．４９％。

In this paper,we present a stochastic model integrating Chinese word segmentation with part-of-speech tagging.We also develop a Chinese lexical analyzer using a two-way searching algorithm which incorporates backward dynamic programming algorithm into A*decode algorithm.The primary experiment proved that the overall accuracy of the proposed analyzer is 98.67% for segmentation and 95.49% for POS tagging respectively.