並列摘要
Micro-bloggers with higher influence are important for information propagation and products recommendation in micro-blogging networks. In this paper, we evaluate micro-blogger's features by follower factor, spreading factor and interaction factor, and integrate these user features with social relationship network structures into a unity graph model. Then we propose an influence analysis algorithm Influence-Rank, which aims to quantify the micro-bloggers' influence. Experimental results show that Influence-Rank can take both the network connectivity and user characteristics into account, dig out influential micro-bloggers who are important for information generation and dissemination. Compared with other methods, Influence- Rank shows better performance in objectivity and effectiveness.