Probabilistic Model Based Large-Scale Social Network Community Discovery Algorithm

In this paper, we propose a novel large-scale social network community discovery algorithm based on probabilistic model to organize users with similar interests into a same group. Firstly, the user community discovery problem is illustrated. The large-scale social network can be regarded as a graph, in which edge represents the relationship between two nodes. Therefore, the user community detection problem can be converted to the graph partition problem. Secondly, our proposed user community discovery algorithm is given. Our algorithm follows an assumption that users of a same community are possible to have same or similar interests. Therefore, the main innovations of our algorithm lie in that the community topics are be represented as multinomial distribution on words, and user interests in different topics obey the probabilistic distribution on community topics. Finally, experiments are conducted to make performance evolution. Experimental results demonstrate that our proposed algorithm can effectively solve the problem of user community detection for the large-scale social network than other methods.