论文部分内容阅读
现在的社交网络不只是人们现实生活社交圈的一个反映,同时也在一定程度上扩展着人们的交际范围,使得用户在社交网上找到更多适合自己的朋友.但是,由于社交网络发展迅速,用户量巨大,对于用户来说,自己从中找到自己的好友是比较困难的,这就需要社交网站向用户提供一个比较好的推荐算法,从而使得网站真正能够改变用户的生活.本文提出的JAFLink(Jaccad-AdamicAdar-Feature)链路加权方法,结合jaccad和AdamicAdar并考虑了社交网络用户属性,构成JAFLink加权方法,用来计算节点对建立链接的可能性.实验表明,JAFLink相比只考虑网络拓扑结构或者用户属性而言更加高效.“,”Social network is not only a reflection of people′s real life,but also to a certain extent,the expansion of the scope of people′s communication,which allow the users find more suitable friends.But because of the rapid development of the social network and amounts of users,it is difficult to users to find new friends.So the users need the help of social network to find new friends,then the social network will give users a better life.This paper attempts to merge the Jaccard Coefficient,the Adamic-Adar and user features to construct an algorithm to calculate the probability about new link between two nodes.The result show that JAFLink is better than the algorithm only consider the user attributes or the network topological structure.