在线社交网络中用户伪装攻击检测方法研究

当前用户伪装攻击检测方法无法适应动态环境,实时性不高;且需要准确的先验知识,检测精度较低。提出一种新的在线社交网络中用户伪装攻击检测方法,介绍了k最邻近节点(KNN)算法的基本思想,给出KNN算法的实现过程。分析了用户伪装攻击检测与分类的关系,确定在线社交网络中用户伪装攻击检测就是对被检测的未知行为进行分类的过程。针对用户行为,将训练集中正常用户行为的邻居进行排列,通过和k相似的邻居的分类标签对新用户行为类别进行判断,从而实现用户伪装攻击检测。实验结果表明,所提方法不仅检测精度高,而且开销小。

Online social network users in disguise attack detection method research

Current user disguised attack detection method cannot adapt to the dynamic environment, real-time is not high, and need accurate a priori knowledge, low accuracy.Put forward a new user as camouflage detection method in online social network, this paper introduces the K most neighboring nodes (KNN) algorithm, the basic idea of the implementation process of KNN algorithm are given.Analysis of the relationship between user disguised attack detection and classification, determine the online social network users in the disguise of attack detection is to detect the unknown behavior classification process.On user behavior, the training focus on normal user behavior of neighbors, with similar k neighbor classification category labels to new users behavior of judgment, so as to realize the user disguised attack detection.The experimental results show that the proposed method is not only high precision, and low overhead.