InfluenceRank: An Improved Online Social Influence Model

Abstract

User influence is a popular research content in online social networks, and it plays an important role in marketing, public opinion management, and network relationships. Traditional research on user influence based on graph structure mainly considers whether users follow each other. However, "zombie fans" can make the influence analysis results inaccurate. Based on the PageRank algorithm, this study proposes a novel model for measuring user influence: InfluenceRank. User behavior and interaction information are introduced into the model through three indicators: activity, interaction and credibility. The experimental results, which are more comprehensive and persuasive, prove that the influence ranking of the InfluenceRank model on Microblog (Chinese Twitter) users is not limited to the number of users' fans.