中文摘要：

提出一种新的朋友推荐方法,该方法同时使用用户兴趣和朋友关系这2种因素来为目标用户推荐朋友,对PageRank算法进行改进,提出一种能同时融合上述2种因素的Topic_Friend_PageRank（TFPR）模型。首先,采用LDA（latent Dirichlet allocation）分析用户发布的消息内容,将用户表示为若干主题上的分布,从而建模用户的兴趣。接下来,使用加权的PageRank算法建模用户在整个链接拓扑中的重要程度和用户之间朋友关系的相似性。最后根据主题感知的PageRank思想,将用户兴趣融入前面提到的加权PageRank中,形成同时融合用户兴趣和朋友关系的TFPR模型。采用新浪微博数据验证所提模型的性能,实验证明该模型能同时得到较高的准确率和召回率。

英文摘要：

A new hybrid approach by incorporatin gusers＇ interests and users＇ friendships together to recommend new friends for target users is proposed. A variation of Page Rank—Topic_Friend_Page Rank（TFPR） is proposed, which can consider user interests and user friends at same time. Firstly, proposed method uses latent Dirichlet allocation（LDA） to model users＇ interests, and weighted-Page Rank algorithm to model users＇ friendship network, and then merge these two factors into TFPR. This hybrid method models users＇ interests and users＇ friendships at the same time, and wedemonstrate the effectiveness of proposed hybrid model by using some social network datasets.