结合信任和用户关系的微博关注推荐算法.

Users followee recommendation plays an important role in information sharing over microblogging platforms. The majority existing followee recommendation schemes adopt either content or background information relevance to measure the user distance, and most use a single factor for followee ranking, there is no classification for user groups, suffering poor performance. Thus this paper proposed a novel followee recommendation algorithm RTLR for non-celebrity that took into user the information sources of both followee relationship and interaction behavior, and established logistic regression model combined with multi-factors. It conducted experiments on real-world datasets collected from Sina Weibo, the largest microblogging system in China. The results demonstrate that the novel scheme effectively improves the accuracy of followee recommendation. [ABSTRACT FROM AUTHOR]