基于多元词组和数据流聚类的热点话题动态发现

Discovering of text's hottopic based on the dynamic phrases and stream data mining

本文主要通过改进的TF-IDF算法和多元词组动态构建来选择特征关键词,并利用Clu Stream数据流聚类方法,实现文本主题的动态发现.实验表明,该方法可以较好地发现海量文本信息中不断变化的主题信息,从而达到推荐关联主题、动态监测舆情等目的.

This paper rapidly find text’s hot topic by the improved TF- IDF algorithm and dynamic phrases that can choose the key words for text. The second,we uses clustering evolving datastreams to get hot topic for the candidate text. Discovering of text topic can detect the most important aspects of the vast information,so as to achieve the monitoring of public opinionrapidly.