摘要
深入挖掘社交网络中传播力较强的个体,并利用其进行产品营销往往会达到事半功倍的效果,影响最大化问题就是在特定社交网络中寻找影响力较大的个体.为了更加准确的评估影响力,本文不仅从节点相似度方面进行改进,而且从信息内容本身出发,基于信息在社交网络中的传播,结合信息词频等信息自身特点来刻画节点的影响力,提出了基于信息词频和节点相似度的影响最大化算法(IMFS,Influence Maximization algorithm based on term Frequency and node Similarity).随后,在真实的社交网络中对该算法进行了实验,并与传统的影响最大化算法对比,实验结果表明由IMFS得到的集合的影响范围大于其他启发式算法的结果,同时算法的运行速度也有相应的提高,说明了本文提出的算法是解决影响最大化问题的有效算法.
Deeply mining the individual who has strong propagation force in social network, and using it to promote products tends to achieve more with less. Influence maximization algorithm aims to solve this problem in the specified social network. In order to esti- mate the influencemore accurate, we not only improve the nodes similarity, but also consider the information content itself. Based on the transmission of information in a social network, combining with the term frequency and other information characteristics to depict the node＇s influence,we proposed a new algorithm based on the diffusion of information,which considered the perspective of term fre- quency and node similarity. We implement this algorithm in the real world data set. Compared to traditional influence maximization al- gorithms, the test result shows that our algorithm is better than other heuristic algorithms. Moreover, our algorithm also runs faster than others. It is verified that our algorithm is an effective solution to influence maximization problem.
出处
《小型微型计算机系统》
CSCD
北大核心
2017年第2期259-263,共5页
Journal of Chinese Computer Systems
基金
国家自然科学基金项目(61502170)资助
关键词
影响最大化
节点相似度
信息词频
EM算法
influence maximization
node similarity
term frequency
EM algorithm
