摘要
在最大匹配法(The Maximum Matching Method)的基础上,提出了一种回溯正向匹配(The Backtracking Matching Method)的中文分词方法。该方法首先对待切文本进行预处理,将文本分割成长度较短的细粒度文本;利用正向匹配、回溯匹配和尾词匹配来有效发现歧义字段;利用长词优先兼顾2词簇的方式对交集型歧义字段进行切分。最后对该算法进行的定性分析说明了该方法的先进性,从随机抽取大量语料实验结果上证明了该方法的有效性。
Chinese word segmentation methods of the backtracking matching method is proposed based on the maximum matching method.The text is pretreatment by the BM in the first,then,the text is cut into shorter lengths granular text;Ambiguity field is found effective by forward matching method,backtracking matching method and last words matching Method;Crossing ambiguity field is cut by l_~ng term priorities and 2-words rules.In the last,the algorithm of the qualitative analysis shows that the method of the advanced nature.The large number of randomly selected language materials being tested and results show that the method is effective.
出处
《计算机工程与应用》
CSCD
北大核心
2009年第22期132-134,共3页
Computer Engineering and Applications
关键词
中文分词
最大匹配法
回溯匹配
交集型歧义
Chinese word segmentation
Maximum Matching（MM） methods
backtracking matching
crossing ambiguity
