基于用户粉丝聚类现象的微博僵尸用户检测

Detecting Zombies in Microblog Based on the Clustering Phenomenon of Fans

随着微博人气的日益高涨,僵尸用户的数量正以惊人的速度增长,虚假导致的微博信任危机严重影响了微博的发展.目前普遍依据关注数、粉丝数、原创和转发信息频率等用户基本属性来判定僵尸粉.然而,微博用户类型纷繁复杂,存在大量的误判和漏判现象.本文通过从用户的粉丝中挖掘凝聚子群,并结合用户的社会网络关系,提出一种基于用户粉丝聚类现象的僵尸粉检测模型.实验结果表明,本模型只需要少量信息就可以有效地对僵尸粉进行检测.

With the growing popularity of microblog, the amount of zombies is increasing rapidly. The crisis of confidence caused by fake seriously affects the development of microblog. At present,most of the technologies of detecting zombies are based on the basic characters of microbloggers such as the followings＇ counts, the followers＇ count,the frequency of writing tweets and retweeting. How- ever, since the kinds of microbloggers are complicated, there are a lot of misjudgments and false negatives. The paper proposes a zom- bies detecting model based on the clustering phenomenon of fans by mining communities from microbloggers＇ fans,and users＇ social networks. The experimental results show that the model can detect zombies effectively with less information.