欢迎访问中国科学院大学学报,今天是

中国科学院大学学报 ›› 2014, Vol. 31 ›› Issue (5): 691-700.DOI: 10.7523/j.issn.2095-6134.2014.05.016

• 计算机科学 • 上一篇    下一篇

基于群体特征的社交僵尸网络检测方法

倪平, 张玉清, 闻观行, 刘奇旭, 范丹   

  1. 中国科学院大学 国家计算机网络入侵防范中心, 北京 101408
  • 收稿日期:2013-10-15 修回日期:2013-12-16 发布日期:2014-09-15
  • 通讯作者: 倪 平,E-mail:nip@nipc.org.cn
  • 基金资助:

    国家自然科学基金(61272481,61303239)和北京市自然科学基金(4122089)资助

Detection of socialbot networks based on population characteristics

NI Ping, ZHANG Yuqing, WEN Guanxing, LIU Qixu, FAN Dan   

  1. National Computer Network Intrusion Protection Center, University of Chinese Academy of Sciences, Beijing 101408, China
  • Received:2013-10-15 Revised:2013-12-16 Published:2014-09-15

摘要:

攻击者通过在社交网络中部署由大量社交僵尸账号组成的社交僵尸网络,对社交网络进行渗透,严重危害了社交网络和用户的信息安全.我们首次提出一种基于群体特征的社交僵尸网络检测方法.提取社交僵尸网络中账号注册时间集中、昵称相似和活跃时间一致3个群体特征,结合数据挖掘算法,设计一种社交僵尸网络的检测方法.在对新浪微博中48万个账号的检测实验中,检测出多个社交僵尸网络,共包含6 899个社交僵尸账号.较低的漏报率和误报率表明该方法对于社交僵尸网络和僵尸账号的检测是可行和有效的.

关键词: 社交僵尸账号, 社交僵尸网络, 社交网络, 数据挖掘

Abstract:

An adversary can infiltrate online social networks (OSNs) on a large scale by deploying socialbot network, which is an army of socialbot accounts. This will endanger the information security of online social network and users. To solve the problem, we propose a detection method based on the population characteristics. We extract the following population characteristics: centralized created time, similar screen names, and coincident active time. On the basis of the extracted charateristics and by using date mining method, the method is proposed to detect socialbots networks. The method is used in a data set of 480 000 users of sina microblog and detects many socialbots networks which include 6 899 socialbots accounts. The low false negative rate and false positive rate indicate that the method is feasible and effective.

Key words: socialbots accounts, socialbot networks, online social networks, data mining

中图分类号: