• 中国计算机学会会刊
  • 中国科技核心期刊
  • 中文核心期刊

J4 ›› 2015, Vol. 37 ›› Issue (11): 2134-2141.

• 论文 • 上一篇    下一篇

基于逆向游走的PageRank社交网络影响力度量算法

郑孝遥,杨文建,鲍煜,罗永龙   

  1. (安徽师范大学数学计算机科学学院,安徽 芜湖 241003)
  • 收稿日期:2015-08-03 修回日期:2015-10-11 出版日期:2015-11-25 发布日期:2015-11-25
  • 基金资助:

    国家自然科学基金资助项目(61370050,61572036);安徽高校自然科学研究资助项目(KJ2015A067,KJ2014A088);芜湖市科技计划重点资助项目(2015cxy10);安徽师范大学校创新基金资助项目(2015cxjj10)

A social networks influence measurement
algorithm based on reverse walk PageRank 

ZHENG Xiaoyao,YANG Wenjian,BAO Yu,LUO Yonglong   

  • Received:2015-08-03 Revised:2015-10-11 Online:2015-11-25 Published:2015-11-25

摘要:

随着社交网络的发展,其节点影响力度量成为一个重要的研究领域。针对传统随机游走PageRank算法精确度不高的问题,提出一种逆向随机游走PageRank算法,该算法采用逆向查找消息传播源的思想,对网络中的每条有向边以概率ε进行逆向随机游走,通过迭代计算出每个节点的PageRank值。实验表明,本文提出的算法较传统的随机游走PageRank算法具有更好的稳定性,并在迭代次数较少时也能保持较高精度。

关键词: PageRank, 随机游走, 社交网络, 影响力度量

Abstract:

(School of Mathematics and Computer Science,Anhui Normal University,Wuhu 241003,China)Abstract:With the development of social networks, the influence measurement of nodes has become an important research area. Aiming at the accuracy problem of the traditional random walk PageRank algorithm, we in this paper propose a reverse random walk PageRank algorithm, which is based on the idea of reversely searching dissemination source, each directed edge starts the random walk with probability ε,and the value of PageRank is calculated by iteration. Experimental evaluation on publicly available datasets demonstrates that our algorithm has the improved stability and higher accuracy when there is less iteration.

Key words: PageRank;random walk;social networks;influence measurement