| 复杂合著网络中的重叠社团发现与可视化 |
| |
| 引用本文： | 谷瑞军,陈圣磊,陈耿,汪加才.复杂合著网络中的重叠社团发现与可视化[J].图书情报工作,2012,56(12):72-59. |
| |
| 作者姓名： | 谷瑞军 陈圣磊 陈耿 汪加才 |
| |
| 作者单位： | 南京审计学院信息科学学院 |
| |
| 基金项目： | 回家自然科学基金项目“基于DM技术的企业舞弊分析的审汁服务系统研究”，旧家社会科学旗金项日“我国反洗钱市计方法研究” |
| |
| 摘 要： | 指出随着互联网的发展和社交网络的广泛应用，学者之间的合作途径越来越多，学者具备多重的社团属性；但是，传统的基于聚类或模块度优化等社团划分方法往往将学者划分到唯一的社团。基于Salton方法构建合著网络，使用加权的链接聚类算法实现社团的聚类划分，该方法将节点间的边作为聚类对象，采用凝聚式层次聚类进行社团划分。因节点属于不同的边，因而可以归属于不同的社团，最终得到的社团可以部分重叠。为检验方法的有效性，使用基于CDPLP的合著网络构建系统获取数据，构建合著网络；然后使用加权的链接聚类进行社团发现和可视化。结果表明，该方法能有效地发现部分重叠的合著社团，且社团的意义比较明确。
|
| 关 键 词： | 重叠社团 链接聚类 合著网络 复杂网络 |
| 收稿时间： | 2011-10-31 |
| 修稿时间： | 2012-02-14 |
Overlapping Community Discovery and Visualization in Complex Co-authorship Network |
| |
| Gu Ruijun,Chen Shenglei,Chen Geng,Wang Jiacai.Overlapping Community Discovery and Visualization in Complex Co-authorship Network[J].Library and Information Service,2012,56(12):72-59. |
| |
| Authors: | Gu Ruijun Chen Shenglei Chen Geng Wang Jiacai |
| |
| Institution: | School of Information Science, Nanjing Audit University, |
| |
| Abstract: | With the development of Internet and SNS, there are so many ways to carry out academic cooperation between scholars that one scholar may belong to several communities. However, every scholar must belong to only one community when using traditional community discovery methods. In this paper, co-authorship network is constructed based Salton relation strength and community discovery in co-authorship network is performed based on weighted link clustering. In this method, weighted link is regarded as clustering object and one node may belong to several clusters. As a result, overlapping co-authorship community can be identified. To validate this method, several co-authorship networks were constructed based on Co-authorship Network Construct System developed by us and then experiments were carried out on co-authorship networks using weighted link clustering. The result shows that our method can find explicit co-authorship communities with partial overlapping and each community is more characteristic compared with others. |
| |
| Keywords: | overlapping eomnmnity link clustering co-authorship network complex network |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 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 |
