Identifying Opinion Leader in the Internet Forum

Abstract

Opinion leader is an authority person who has great influence in BBS. Their linguistic behavior has a huge impact on net citizen's behavior and thought. In this paper, we propose an algorithm called OLRA (Opinion Leader PageRank Algorithm) based on topic-field to identify opinion leaders in the Internet forum. In the algorithm the closeness degree factor and sentiment analyses are taken into consideration. Meanwhile, these two authority values are defined as the weight of links among users. The data is collected from a number of posts on Tianya forum which is a famous forum in China. In the experiment, the algorithm is compared with Interest-based PageRank algorithm, online time Algorithm, and Experience-based Algorithm, the result shows that the OLRA algorithm can identify opinion leaders than others in the Internet forum effectively.