基于多维社交关系的在线社交网络链路预测研究  

Research on Online Social Networks Link Prediction Based on Multidimensional Social Relations

作　　者：何军[1,2] 刘业政[1] He Jun;Liu Yezheng(School of Management,Hefei University of Technology,Hefei 230009,China;School of Business,Anhui University,Hefei 230601,China)

机构地区：[1]合肥工业大学管理学院,安徽合肥230009 [2]安徽大学商学院,安徽合肥230601

出　　处：《现代情报》2017年第7期41-46,115,共7页Journal of Modern Information

基　　金：国家973计划课题"社交网络分析与网络信息传播的基础研究"(项目编号:2013CB329603);国家自然科学基金项目"基于模体挖掘面向在线社交网络中虚拟社区的群推荐系统研究"(项目编号:71371062);安徽省教育厅人文社科重点项目"在线社交网络社区形成机制对企业社会化营销策略的影响研究"(项目编号SK2015A234)

摘　　要：能有效预测在线社交网络中人的行为有较高的应用价值,对当前在线社交网络存在多种交互方式的现象进行分析,以此构建一种多维社交关系的在线社交网络链路预测的方法。基于社会影响和选择两种机制构建在线社交网络多维网络结构,综合各个维度节点和连边的拓扑结构,实现多维社交网络预测矩阵的集结,在降维后的网络上进行链路预测。利用百度贴吧数据集实验验证,从AUC和精确度两方面评价模型,结果表明本文提出的多维预测方法是有效的。It's valuable to effectively predict the behavior of people in online social networks.Based on the analysis of the phenomenon of many kinds of interaction in online social network,this paper built an online social network link prediction method based on multidimensional social relationship.The paper constructed a multidimensional online social networks structure based on social influence and selection mechanism.Took into account the topology of the nodes and edges about each dimension,to realize the aggregation of multidimensional social networks prediction matrix.Finally carried out link prediction on the reduced dimension networks.Th is paper used Baidu PostBar data set to evaluate the model from both AUC and precision,the result showed that the method proposed in this paper was eff ective.

关 键 词：多维网络 在线社交网络 链路预测 百度贴吧 

分 类 号：O157.5[理学—数学]