Identifying User Identity across Social Network Sites based on Overlapping Relationship and Social Interaction

Abstract

Most people have identities on multiple social network sites (SNSs) simultaneously to meet their diverse needs for social media use. User identity identification across SNSs has been a significant research focus in recent years as it is important for comprehending users' online behaviors, conducting precise recommendations, detecting cross-site identity attacks, etc. In this paper, we propose a new method for user identity identification across SNSs by utilizing users' overlapping relationships and corresponding social interactions among different SNSs. In our method, the overlapping relationships are leveraged to generate user identity candidates for each identity to be identified, and the user identity matching procedure is performed based on username, user content, social network as well as social interaction. We conduct experiments by setting Weibo as source SNS and Douban as target SNS as well as Douban as source SNS and Weibo as target SNS respectively. The experimental results indicate our method's better performance on precision, recall, F1-score and accuracy.