一种改进的新浪微博好友推荐算法

目前新浪微博的好友推荐机制存在一些缺点,通过研究微博社区结构和权威用户对好友推荐的影响,提出了一种改进的基于新浪微博的好友推荐算法。在同一微博社区中,通过查找权威用户,并结合用户之间的兴趣相似度和信任度进行好友推荐,推荐过程中两次计算了用户的兴趣相似度并引入用户间信任度传播模型。选取微博社区中目标用户进行实验的结果表明,权威用户在好友推荐中起了重要作用,提高了好友推荐的效果。同时通过将该算法推荐的好友列表和原新浪微博推荐的好友列表作对比,实验表明该算法具有较好的推荐效果。

A Modified Friend Recommending Algorithm Based on Sina Microblogging

There are some deficiencies in the current Sina microblogging friend recommending mechanism.By studying the influence of microblogging community structure and authority users on the friend recommendation,this paper proposed a modified friend recommending algorithm based on Sina Microblogging.In the same microblogging community,conducting friend recommendation by searching the authority should be combined with interest similarity and trust degree between users.In the friend recommending process,the algorithm calculated the interest similarity between users twice and introduces user trust degree propagation model.The results of selecting target users in real microblogging community to carry on the experiments show that the authority users play an important role in friend recommendation and can improve the effect of friend recommendation.At the same time,the results of contrasting the Sina microblogging friend recommending list and the proposed algorithm friend recommending list show that the proposed algorithm has a better recommending effection.