【摘要】 近些年,微博、博客、人人网、豆瓣等各种社交网络的发展,引发了学术界对社会网络、虚拟社区的关注.本论文提出了一种基于用户转发的User-Behavior Rank(UBR)算法,力求在推荐过程执行之前,预先选出微博社区中具有权威性的博主,提高推荐新颖性.文中的基于转发行为的User-Behavior Rank是对传统的Page Rank(PR)的改进,以博主之间的关注关系为边,基于其发微博的行为以及其粉丝对微博的转发,计算各自的UBR值,取UBR值较大的博主作为推荐集合.与Page Rank相比,UBR能够选出更加符合社会实际的有影响力的用户集.更多还原

【Abstract】 In recent years,with the rapid development of a variety of social networking platforms,such as micro-blog,blog,Renren,Douban,and so on,the research on the recommendation of social network and virtual community has attracted more and more attention. In this paper,we propose a User-Behavior Rank algorithm(UBR)based on user forwarding behavior,and strive to select the influential users in the micro-blog community before the implementation of the recommendation algorithm. The User-Behavior Rank algorithm based on user forwarding behavior is an improvement of the traditional PageRank(PR)algorithm:the concernrelationship between users is as edge,basing on user’s micro-blog behavior and fan’s micro-blog forwarding to compute the UBR value of network nodes.And taking UBR value ranking in the top users as the recommended object. Compared with the traditional PageRank algorithm,it is possible to select an influential user set that is more in line with the social reality.更多还原