关联规则在汉语词属性中的研究

摘要

在语音识别和合成中文本分析是很重要的，文本分词是文本分析正确与否的基础。目前语音合成系统中的分词一般是基于词典分析建立的，对于多音字易产生错误。文章基于数据挖掘中的关联规则的发现方法对文本分词中词语的关联关系进行研究，通过文本数据的文本特征及语音特征描述的有机结合，获取词语自身属性的关联关系，最后进行了实例测评。%Text Analysis is very important in Speech Recognition and Synthesis. Word Segmentation is on the basis of text analysis. The models in most speech synthesis systems that are now being used are constructed by dictionary,qualitatively and with high errors. This paper applies data mining approach to finding association rules from text data.Through analyzing text attributes to decrease word segmentation errors. The Rules can improve segmentation precision.