基于信任社交圈的好友推荐算法

Friend Recommendation Algorithm Based on Trusted Social Circle

为体现出在线社交网络中好友推荐时的用户倾向性,并且更真实反映现实生活中人与人之间的关系强度,将用户间的好友关系强度定义为信任度引入到相邻边拓扑信息相似性计算中,并结合用户兴趣模型导出用户社交圈,提出一种好友推荐算法。针对用户社交圈中未考虑好友间信任度的情况,将信任度融入到社交圈重叠程度计算中为用户发现其潜在的好友提供建议。实验结果表明,与基于社交圈相似性和公共朋友相似性的好友推荐算法相比,该算法具有更高的好友推荐准确率。

In order to reflect the tendency of users in the recommendation of friends in online social network, also to more really reflect the relationship strength between people in real life, the relationship between users is defined as the credibility and applied into the similarity calculation of topological information of adjacent edges, which is combined with user interest model to derive user social circle, then proposes a friend recommendation algorithm. Aiming at the situations of not taking into account the credibility between friends, the credibility is applied into the calculation method of the overlap degree of social circle, so as to provide suggestions for the users to find their potential friends. Experimental results show that the proposed algorithm has higher accuracy of friend recommendation, compared with friend recommended algorithms based on the similarity of social circle and the similarity of common friends.