首页 | 官方网站   微博 | 高级检索  
     

一种融合节点属性信息的社会网络链接预测方法
引用本文:张昱,高克宁,于戈.一种融合节点属性信息的社会网络链接预测方法[J].计算机科学,2018,45(6):41-45.
作者姓名:张昱  高克宁  于戈
作者单位:东北大学计算机科学与工程学院 沈阳110819,东北大学计算机科学与工程学院 沈阳110819,东北大学计算机科学与工程学院 沈阳110819
基金项目:本文受教育部基本科研业务费项目青年教师科研启动基金(N151603001),辽宁省科技攻关项目博士启动基金(201601026)资助
摘    要:随着大规模社会网络的发展,链接预测成为了一个重要的研究课题。研究了在社会网络中融合节点属性信息进行链接预测,在传统的社会-属性网络图模型的基础上,将节点属性的类别这一重要参量加入到网络构建中。基于此,提出了一系列为网络中不同类型的连边分配边权重的方法,最后通过随机游走的方法进行网络链接的预测。实验表明,所提链接预测方法相比同类方法有明显的效果提升。

关 键 词:链接预测  社会网络  社会-属性网络  社会节点  属性节点
收稿时间:2017/3/11 0:00:00
修稿时间:2017/6/18 0:00:00

Method of Link Prediction in Social Networks Using Node Attribute Information
ZHANG Yu,GAO Ke-ning and YU Ge.Method of Link Prediction in Social Networks Using Node Attribute Information[J].Computer Science,2018,45(6):41-45.
Authors:ZHANG Yu  GAO Ke-ning and YU Ge
Affiliation:School of Computer Science and Engineering,Northeastern University,Shenyang 110819,China,School of Computer Science and Engineering,Northeastern University,Shenyang 110819,China and School of Computer Science and Engineering,Northeastern University,Shenyang 110819,China
Abstract:With the development of large social networks,link prediction has become an important research subject.The link prediction problem in social networks using rich node attribute information was studied in this paper.Based on attribute-augmented social network model,which means rebuilding an augmented network by adding additional nodes with each node corresponding to an attribute,called social-attribute network,the classification of node attributes was added to the model as an important parameter.Several methods of assigning weights for different kinds of links were proposed.Then a random walk method was used for link prediction in the network.Experimental results reveal that this method has better performance compared with other similar methods.
Keywords:Link prediction  Social network  Social-attribute network  Social node  Attribute node
点击此处可从《计算机科学》下载全文
设为首页 | 免责声明 | 关于勤云 | 加入收藏

Copyright©北京勤云科技发展有限公司    京ICP备09084417号-23

京公网安备 11010802026262号