Opinion Leader Mining of Social Network Combined with Hierarchical Sentiment Analysis

As social networks play more and more important roles in daily life, some malicious people use them to disseminate disinformation. For mining opinion leaders, we propose a method combined with sentiment analysis to find suspicious as well as dominant users. This mining method takes the tree to interpret the users’ interaction in social network for calculating users’ influence score. Besides, a hierarchical algorithm is used for emotional analysis. In order to search for users who probably against the national security, experiments on the dataset from Sina Weibo show our procedure has an advantage over traditional methods.