| 面向大规模多维社会网络的社区发现研究 |
| |
| 引用本文： | 吴小兰,章成志.面向大规模多维社会网络的社区发现研究[J].图书情报工作,2014,58(16):122-130. |
| |
| 作者姓名： | 吴小兰 章成志 |
| |
| 作者单位： | 1. 徽财经大学管理科学与工程学院;
2. 南京理工大学信息管理系 |
| |
| 基金项目： | 本文系国家社会科学基金项目“在线社交网络中基于用户的知识组织模式研究”（项目编号：14BTQ033）和安徽财经大学2014年度校级科研项目“大尺度在线社会网络的社区发现及其应用研究”（项目编号：ACKY1428）研究成果之一。 |
| |
| 摘 要： | 认为处于多维社会网络中的用户会表现出多种行为取向和兴趣爱好，单独使用多维网络中的一维很难进行有效的社区发现。为了有效解决以上问题，首先基于用户关系紧密度将社交媒体中有向网转化为无向带权网，并将所有一维社交网络进行集成；然后利用SSN-LDA对社交用户进行隐含社区建模，以根据用户-隐含社区概率分布计算用户相似度；最后使用二分K均值进行用户社区划分。在真实科学网博客上进行试验，结果表明该方法能较好地进行用户社区划分。
|
| 关 键 词： | 多维网络 社区发现 主题模型 大规模网络 |
| 收稿时间： | 2014-05-27 |
Community Detection for Large-scale Multi-dimensional Network |
| |
| Wu Xiaolan,Zhang Chengzhi.Community Detection for Large-scale Multi-dimensional Network[J].Library and Information Service,2014,58(16):122-130. |
| |
| Authors: | Wu Xiaolan Zhang Chengzhi |
| |
| Institution: | 1. School of Management Science and Engineering, Anhui University of Finance and Economics, Bengbu 233030;
2. Department of Information Management, Nanjing University of Science and Technology, Nanjing 210094 |
| |
| Abstract: | Users in multi-dimensional network usually show a variety of behaviors and interests, so it is hard to find effective communities by using only one dimension. In order to effectively solve the above problem, this paper firstly maps directed networks into undirected weighted networks based on user relationship strength, and then integrates all the networks. Secondly, this paper models hidden community by using SSN-LDA, and calculates the users' similarity by user-community probability distribution matrix. At last, Bisecting K-Means is used to detect community of users. Through the experiments on real Science blog, the result shows that this method can get more accurate user community. |
| |
| Keywords: | multi-dimensional network community detection topic model large-scale network |
|
| 点击此处可从《图书情报工作》浏览原始摘要信息 |
| 点击此处可从《图书情报工作》下载免费的PDF全文 |
