Social User Influence Measurement Algorithm Based on Multi-perspectives Analysis

With the rapid development of social networks, user influence plays an increasingly vital role in commodity marketing and public opinion monitoring. Therefore, analyzing and identifying users with high influence has become a hot research topic. Aiming at the lack of comprehensive analysis in existing influence measurement methods, a novel user influence ranking method ASIRank (Actively Strength and Interest based Rank) is proposed to measure influence of social users. First, in order to fully consider the contribution of different relationships to user's influence, the user topological networks are constructed from multiple perspectives such as follow, retweet, comment and mention. Secondly, for the sake of quantifying influence more accurately, factors such as user activeness, interaction intensity and interest preferences are considered in different relational networks. Then, user's influence from different perspectives is integrated to obtain the influence ranking of users in the global scope. Through experiments on real microblog data set, the results show that the proposed method is superior to existing methods in terms of accuracy, correlation and coverage.