查看文章

There are vigorous developments of social network which affect out life greatly. User influence is an important reason to promote the interaction in social network. When we analyze user influence, single value can't indicate the user influence in different domains. This paper puts forward the design of User Classification PageRank (UCPR) to solve this problem. Firstly, we classify users according to the content which they forwarded. Then, we use space mapping to set up several subnet. Finally, we analyze user influence in every specific subnet by Domain Mapped Network (DMN) which is based on PageRank algorithm and we improve this algorithm to analyze the user influence in different domains. Through the work of this paper, we used a vector to present user influence rather than a single number and we test and verified the long-tailed distributions of social network in experiments.