基于话题聚类及情感强度的中文微博舆情分析

Analysis of the Public Opinion of Chinese Microblog Based on Topic Clustering and Emotional Intensity

文章通过话题聚类及情感强度分析中文微博舆情,实现对微博热点问题的预测,有利于公众舆情引导。首先充分考虑微博短文本的特点,在特征值提取基础上克服了微博短文本易发生"文本漂移"的缺点,并根据微博高频词对微博进行排序实现微博的快速聚类,接着从主观和客观两方面对热点话题的情感强度进行了分析,基于灰色模型跟踪并预测公众情感变化倾向。实验结果表明,本文提出的基于话题聚类及情感强度的中文微博舆情分析方法具有一定的可行性。

Based on the topic clustering and emotional intensity, this paper analyzes the public opinion of Chinese microblog and realizes the prediction of hot topics, which is benefit for the guidance of public opinion. Firstly, the paper fully considers the characteristics of short texts of microblog. Then, on the basis of feature extraction, the paper overcomes the shortcomings that short texts of microblog easily occur in the ＂text drift＂ . According to the high-frequency words of microblog, the paper sorts the microb- log in order to achieve the fast clustering. The paper analyzes the emotional intensity of hot topics from the aspects of subjective and objective. Finally, the paper follows and predicts the tendency of public emotion based on the grey model. The result shows that the analysis method of public opinion of Chinese microblog based on topic clustering and emotional intensity has certain feasibility.