Exploit of online social networks with community-based graph semi-supervised learning

摘要: With the rapid growth of Internet, more and people interact with their friends in online social networks like Facebook1. Currently, privacy issue becomes a hot dynamic research topic. Though some protecting strategies are implemented, they not stringent enough. Recently, Semi-Supervised Learning (SSL), which has advantage utilizing unlabeled data to achieve better performance, attracts much attention from web community. By large number websites, SSL can effectively infer hidden or sensitive information on Internet. Furthermore, graph-based is suitable for modeling real-world objects graph characteristics, networks. Thus, we propose novel Community-based Graph (CG) model that be applied exploit security issues networks, then provide two consistent algorithms satisfying distinct needs. In order evaluate effectiveness this model, conduct series experiments synthetic StudiVZ2 Facebook. Experimental results demonstrate our approach accurately confidently predict users, comparing previous models.