利用行为模式对用户分类是一个非常新颖的问题,目前基于行为模式的社会网络用户聚类相关研究较少.在社会网络用户的交互行为的基础上构建用户行为马尔可夫模型,并采用一步转移矩阵、n步转移矩阵和收敛马尔可夫分布表达用户行为马尔可夫模型,提出相应的实现算法.基于谱聚类的思想,提出基于行为模式的社会网络用户谱聚类算法,基于行为模式的社会网络用户谱聚类能够发现行为模式相似程度较高的用户群.在人人网和脸谱网上进行了大量实验,实验结果表明本文方法对用户分类效果优于k最近邻算法.在大量用户聚类上,提出方法的聚类结果在聚类密集性和类别差异度上也都优于K最近邻算法. It is a very novel problem to classify users by using behavioral patterns.At present, there are few researches on clustering of social network users based on behavioral patterns.Based on the interaction of social network users, a Markov model of user behavior is constructed and a one-step transfer Matrix, n-step transfer matrix and convergent Markov distribution to express user behavior Markov model, and put forward the corresponding implementation algorithm.According to the idea of ​​spectral clustering, this paper proposes a social network user spectrum clustering algorithm based on behavioral pattern, based on behavioral model The social network users’ spectral clustering can find out the user groups whose behavior patterns are similar to each other.A large number of experiments have been conducted on Renren and Facebook, and the experimental results show that this method is better than k nearest neighbors in the classification of users. On the class, the proposed clustering results are also superior to the K nearest neighbor algorithm in terms of clustering intensity and category diversity.