基于规则挖掘和Naive Bayes方法的组合型歧义字段切分

组合型歧义字段切分是中文自动分词的难点之一。在对现有方法进行深入分析的基础上，提出了一种新的切分算法。该算法自动从训练语料中挖掘词语搭配规则和语法规则，基于这些规则和Naive Bayes模型综合决策进行组合型歧义字段切分。充分的实验表明，相对于文献中的研究结果，该算法对组合型歧义字段切分的准确率提高了大约8%。

Resolving combinational ambiguity in Chinese word segmentation based on rule mining and Naive Bayes method

Combinational ambiguity is one of the most difficult problems for Chinese word segmentation. After in-depth analysis of the other algorithms in literature, the paper proposed a new segmentation algorithm. The algorithm automatically mined word collocation rules and grammar rules from training corpus, and then made integrated decisions to resolve combinational ambiguity based on the mined rules and Naive Bayes method. Extensive experiments show that the proposed algorithm obtains an accuracy increase of 8% against the related works.