摘要
提出一种基于友好度的影响力最大化算法。利用社交网络中用户行为信息计算友好度并构建友好关系网络,通过友好度调整节点的扩散概率,同时启发式地选取友好度积累值较高的节点作为启发节点以此提升算法的实现效率,最后采用CELF++算法计算影响力最大的节点集。实验结果表明,该算法与CELF++算法、TIM算法和DegreeDiscount算法相比,在获得较好的扩散效果的同时亦将执行时间有效的控制在一定范围内。
A new influence maximization algorithm based on friendliness is proposed in this paper and named as Friendliness-based Influence Maximization Method(FIMM).FIMM uses behavior and interaction information of nodes in social network to compute friendliness and to build friendly relationship network.FIMM also adjusts the diffusion probability of nodes through the friendliness and improves the efficiency of implementation of the algorithm by selecting nodes.These nodes have higher accumulation of friendliness and have the most potential influence heuristically based on friendly relationship network.Experiment results show that FIMM can get better diffusion effect than CELF+ +,TIM and DegreeDiscount.FIMM can also limit implement time effectively in a smaller range,which means that using friendliness and selecting nodes heuristically can optimize influence maximization algorithm.
出处
《西安邮电大学学报》
2016年第6期118-122,共5页
Journal of Xi’an University of Posts and Telecommunications
基金
国家自然科学基金资助项目(71501156)
中国博士后基金资助项目(2014M560796)
陕西省教育厅科研计划资助项目(15JK1679)
西安邮电大学创新基金资助项目(114-602080048)
关键词
影响力最大化
友好关系
社交网络
社会计算
influence maximization
friendliness
social network
social computing