De-anonymizing social networks: Using user interest as a side-channel

Social networks, such as Twitter, Instagram, Facebook, and Weibo, have covered a wide range of population throughout the world. People use multiple social networks to enjoy various services as well as share their personal information according to different privacy level. Unlike Facebook and Wechat, in which social connection is more like acquaintance, social networks like Twitter, Instagram, and Weibo represent the social networks in which social link represents interest rather than acquaintance. According to this observation, we propose a community detection method based on interest group, then apply de-anonymization algorithm based on this community. Our experiment shows that this achieves better accuracy than existing de-anonymization algorithm. We conduct several experiments to demonstrate the effect of parameters used in the de-anonymization algorithm.