The theme of the traditional analysis method of hot topic of information mining,on one hand recognition no in-formation on hot topics,wil lead to a low efficiency,on the other hand,too much text topic analysis,the efficiency is too low. Aiming at these problems,based on the Native Bayes classification algorithm,propose a classification algorithm for the social network characteristics of news text,in order to improve the effect of clustering.%随着社交网络的迅速发展,热点话题的提取是目前社交网络中的热门研究方向之一。传统的主题分析方法对消息文本进行热点话题挖掘,一方面识别不出热点话题的相关信息,会导致准确率比较低;另一方面文本太多,使得主题分析效率太低。针对这些问题,在朴素贝叶斯分类算法的基础上,提出一种适合社交网络消息文本特点的分类算法,从而提高聚类的效果。最后,通过实验验证改进后算法的有效性。
展开▼