点击次数：

论文类型： 会议论文

发表时间： 2017-07-06

收录刊物： EI

卷号： 566

页面范围： 327-336

摘要： Nowadays, online social networks (OSNs) play an important role in our daily lives. And it is very common for a person to have many profiles in different OSNs. However, different profiles in different OSNs of the same person are isolated from each other. User Identity Resolution (UIR) is the problem to recognize the same person in different OSNs. Most methods are mainly concerned with the profile attributes and they just use the information of profiles. In this paper, we propose a new algorithm, called Identity Matching based on Propagation of anchor links (IMP) which fully combines the profile attributes, the linkage information and the social actions, and solves the problem by expanding the anchor links (seed account pairs that belongs to the same user). In the IMP algorithm, we use the information of the nodes surrounding the anchor nodes and identify new links. As the spread of the anchor nodes, we can iteratively find more and more links. We conduct extensive experiments on Twitter and Facebook to evaluate our algorithm and the results show that our algorithm significantly improves the matching results and outperforms the baseline algorithms. © Springer International Publishing AG 2018.