基于主题划分的移动社交网络关键位置发现研究

Abstract

目前对关键位置的研究大多侧重于整个网络的拓扑结构,而忽视了节点本身被赋予的社会属性,无法适用于体现用户当前偏好意愿的在线查询。引入主题划分的概念; 来描述节点所具备的不同社会属性。系统分为离线处理和在线查询两个模块,离线环境下主要致力于构建基于主题划分的图网络,以及计算最大路径、最大转移概率; 、影响力上界;; 在线模块根据既有的离线处理,通过加入含有主题划分的查询,将满足条件的关键节点找出,更符合实际需求。实验表明,该系统能够在离线处理后对节点的影响力; 作出正确的评估,在准确率和召回率方面均优于传统的位置推荐算法。

Current research on key location discovery is mostly focused on the; topology of the entire network,ignoring the social attributes.; Therefore,the calculations can not meet the current demand. This paper; introduced the concept of the topic and described the different social; attributes of the nodes.The entire system was divided into two modules:; offline processing and online query processing. The offline module was; focused on building a topic-based network,computing the maximum path,the; maximum migration probability,and the key influence upper bound; calculation. Online module used the query containing topics, to make the; processing more in line with actual demand. Experimental results show; that the system can discover the key node set within an acceptable time; frame and ensure precision and recall rate better than traditional; algorithm.