摘要
社会网络影响最大化问题是当前的研究热点之一。针对SI(Susceptible-Infected)信息传播模型未考虑节点间亲密关系对信息传播的影响,提出一种ESI(Extended Susceptible-Infected)信息传播模型。为避免由于挖掘的初始节点之间的距离选择不当,陷入局部最优影响力,提出一种新的启发式算法——核重构算法CRA(Core Reconstitutions Algorithm)。该算法引入了k阶核心集和重合率的概念,通过重合率合理控制初始节点的影响范围,依次找出影响力最优的节点。基于新浪微博的实验表明,ESI传播模型优于SI传播模型,CRA算法比现有启发式算法具有更优的全局影响效果。
The issue of maximizing social network influence is one of the current research hotspots. In view of SI(Susceptible-Infected) information propagation model does not consider the impact of intimacy between nodes on information dissemination,an ESI(Extended Susceptible-Infected) information propagation model was proposed. In order to avoid the improper selection of the distance between the initial nodes of the mining,and to fall into the local optimal influence,a new heuristic algorithm—Core Reconstitutions Algorithm(CRA) was proposed. The algorithm introduced the concept of k-order core set and coincidence rate. Through the coincidence rate,the initial node 's influence range was reasonably controlled,and the most influential nodes were found in turn. Experiments based on Sina Weibo show that the ESI propagation model is superior to the SI propagation model,and the CRA algorithm has better global effect than the existing heuristic algorithm.
作者
刘钰峰
郅欢欢
周喻鑫
Liu Yufeng;Zhi Huanhuan;Zhou Yuxin(College of Information Science and Engineering, Hunan University, Changsha 410082, Hunan, China)
出处
《计算机应用与软件》
北大核心
2018年第6期279-285,共7页
Computer Applications and Software
关键词
社交网络
信息传播模型
k阶核心集
重合率
CRA
Social network
Information propagation model
K-order core set
Coincidence rate CRA