社会网络中基于主题的影响力最大化算法  

Topic-based influence maximization algorithm on social networks

作　　者：朱玉婷[1] 李雷[1] 施化吉[1] 周从华[1] 施磊磊[1] 徐慧 Zhu Yuting Li Lei Shi Huaji Zhou Conghua Shi Leilei Xu Hui(School of Computer Science ＆ Telecommunication Engineering, Jiangsu University, Zhenjiang Jiangsu 212013, China Daqo Group Company, Yangzhong Jiangsu 212211, China)

机构地区：[1]江苏大学计算机科学与通信工程学院,江苏镇江212013 [2]大全集团,江苏扬中212211

出　　处：《计算机应用研究》2016年第12期3611-3614,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(71271117);江苏省科技支撑计划资助项目(BE2011156)

摘　　要：为了解决现有的影响力最大化研究没有充分考虑主题对影响力节点挖掘的影响而导致特定主题下节点集合的影响范围不大这一问题,提出了一种社会网络中基于主题的影响力最大化算法TIM。该算法首先根据主题敏感阈值对初始节点集进行预处理,剔除干扰节点,再在新的节点集合上分两个阶段进行节点挖掘。第一阶段挖掘主题权威性大的节点,第二阶段挖掘主题影响增量最大的节点,最后综合两个阶段的节点作为结果集并进行实验验证。实验结果表明,相比其他算法,TIM算法挖掘的节点集合在特定主题下的影响范围更大,时间复杂度更低。To solve the problem that recent researches of influence maximization haven ＇t fully considered that topic has an impact on influential nodes mining, which lead to low influence scope under specific topic, this paper proposed a topic-based influence maximization algorithm （ TIM）. This algorithm first pretreated the initial node set according to topic sensitive threshold, and removed the interference nodes, then mined nodes in two stages on the new node set. In the first stage, it mined the nodes with high topic authority; in the second stage, it mined the nodes with the biggest topic influence increment. At last, it combined the two stages of the node as a resuh set and made an experimental verification. The experimental results indicate that the nodes set mined by the proposed algorithm improve the influence scope under specific topic and the algorithm cost less time compare to other influence maximization algorithms.

关 键 词：社会网络 影响力最大化 主题 节点挖掘 节点集 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]