基于影响路径的个性化影响最大化算法  

A personalized influence maximization algorithm based on influence path

作　　者：杨书新[1] 王希[1] 彭秋英 

机构地区：[1]江西理工大学信息工程大学,江西赣州341000

出　　处：《计算机工程与科学》2016年第6期1128-1134,共7页Computer Engineering & Science

基　　金：国家自然科学基金(41362015);江西省科技厅青年科学基金(20122BAB211035);江西省教育厅科技项目(GJJ14431;GJJ14432;GJJ14458)

摘　　要：个性化影响最大化问题是近年来社交网络影响最大化问题研究领域一个较新的分支,其现有解决方案普遍建立在网络边影响传播强度一致的假设下,该假设对于真实社交网络缺乏普遍适用性。为此基于独立级联模型,提出最大影响路径算法(MIPA)。该算法通过三个阶段来求解个性化影响最大化问题,首先将边影响强度作对数转换以获得最大影响路径,从而计算网络节点对目标节点的邻居节点的影响;然后利用多条经过目标节点邻居的最大影响路径联合计算目标节点受到的影响强度;最后选择Top-k节点作为种子节点,从而摆脱边影响强度的一致性约束,获取高质量的种子集。在不同的真实社交网络数据集上进行的对比实验验证了算法的有效性。Personalized influence maximization in social network has become a new branch of influ ence maximization study in recent years. Different from existing research that assumes equal propagating strengths of social network edges, our work aims to find out the top-k most influential nodes for the target user without inappropriate assumption. We propose a maximized-influence-path algorithm （MIPA） based on the independent cascade model. It solves the problem through three stages. Firstly, to com- pute the propagating strengths from the nodes of social network to the neighbors of the target node, the strengths of edges are transformed into its logarithmic form for getting the maximized influence paths. Secondly, the strength of maximized influence paths which pass through different neighbors with the same source nodes are consolidated to calculate the node＇s propagating strength on the target node. Finally, the seed set with high propagating strength is found out by selecting the top-k nodes. We testify the algorithm on several real world social networks. Experimental results validate the proposed algo- rithm.

关 键 词：社交网络 个性化 影响最大化 特定用户 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]