移动社会网络中基于投票的影响力节点发现

摘要

Influence maximization lies in identifying some seed nodes in a network and maximize the influence in the network. Based on voting model, this paper studies the influence maximi-zation in mobile communication. Firstly, a social relation graph based on mobile communica-tion data is constructed. Secondly, every user votes for its intimate neighbor. As each user has only one ballot and yet has more intimate neighbors, we sort these intimate neighbors according to their liveness, and vote for the most active one. Finally, we sort users according to their bal-lots acquired and return the users on the top of the list. Experiments on real mobile communi-cation dataset validates the effectiveness of the algorithm proposed.%影响力最大化问题是发现网络中的若干节点作为种子节点，使得影响力在网络中的传播最大化。基于投票模型，研究了移动通信网络中的影响力最大化问题。首先，根据用户以及用户间的短消息交互信息生成一个社会关系图。其次，每个用户对其最亲密的邻居进行投票。由于每个用户只能投一票，而其最亲密的邻居可能有多个，按照用户的活跃度对这些亲密的邻居进行排序，并且只对最活跃的邻居进行投票。最后，按照用户得票数对影响力进行排序，并返回排名靠前的若干个。通过对真实的移动通信数据集进行分析验证本文算法的有效性。