The Link Prediction Problem for Social Networks

社交网络的链接预测问题

Given a snapshot of a social network, can we infer which new  interactions among its members are likely to occur in the near future? We  formalize this question as the link prediction problem, and develop  approaches to link prediction based on measures for analyzing the “proximity”  of nodes in a network. Experiments on large co-authorship networks suggest  that information about future interactions can be extracted from network  topology alone, and that fairly subtle measures for detecting node proximity  can outperform more direct measures.

给定一个社交网络的快照，我们可以推断这些用户中最近会有哪些新的交互发生么？我们将这些问题定义为链接预测问题，提出了基于分析节点“接近性”的方法。大型论文合作网络的实验证明，未来交互的信息可以仅通过网络拓扑结构就可以抽取，适当地微妙的方法来探测节点的接近性比直接的方法好。

这篇文章应该算是链接预测的开山之作。定义了链接预测的核心内容——预测近期节点之间的交互关系。然后作者提出了一个重要论断，至少从社交网络上看，链接预测只要拓扑结构就可以了。

这是03年CIKM的文章，总体来讲新在所提出的问题——社交网络的链接预测。