在线社会网络中一种基于时空相似性的多源点定位方法

Detecting Multi-Sources Based on Similarity Between Time and Space in Social Networks

针对在线社会网络中多源点定位问题,当源点数量不确定时,定位准确率有待提高。采用设置探查节点的方式,基于探查节点接收时刻序列分布与网络空间结构具有相似性的特点,提出一种基于时空相似性的多源点定位方法。首先,分析多次接收信息节点的源点指向性,采用重启式随机游走算法确定源点备选集;然后,以非多次接收信息节点与备选源点的时空相似性为基础,将定位问题转化为聚类问题;最后,采用改进的近邻传播算法确定源点数量和位置。实验分析表明,相对于其他算法,该算法可提升源点定位的准确性,减少误差跳数。

The accuracy of locating uncertain number of sources is to be promoted in social network. In this paper we study the diffusion sources locating problem by learning from information diffusion data collected from only a small subset of network nodes. A method for detecting multi-sources based on similarity between time and space in social network (DeMSS) is proposed . First, the directivity of detecting nodes which receive more than one message is analyzed, and potential source nodes set is selected by random walk with restart. Then, according to the similarity between potential source nodes and detecting nodes which receive no more than one message, sources locating is converted to clustering. Finally the sources are located using affinity propagation algorithm. We test and compare the DeMSS using real-world networks to demonstrate its better performance.