文献合作者网络中的拓扑特征与社团结构
首发时间:2010-12-31
摘要:在文献分析领域,随着数据规模的日益扩大,传统意义上的数据分析已经变得越来越难以实现。同时随着网络科学的发展,基于链接的数据分析技术越来越受到人们的重视。本文给出了DBLP作者合作网络的一个分析范例。首先,我们统计了该网络的一些基本统计特征,如度分布,边的权重,以及聚类系数的分布等,并通过可视化技术分析了该网络中的高产作者、作者重名现象以及频繁合作关系等。通过以上的分析,本文列举了一些有趣的特征,如:合作者数量前几名作者均是英文名较短的中国作者,这些作者节点的个人中心网络往往比较稀疏,可能说明在合作网络中存在着明显的重名现象;对比物理与生物作者合作网络,该计算机科学领域的合作网络的平均合作人数较小。在过滤掉网络中的仅出现一次的较不稳定的合作关系后,本文通过使用社团发现算法,对网络进行学术团队划分,找出了社团结构分布特征并提供了基于学术团队的宏观可视化展示,最后本文进一步深入分析了不同团队的内部结构与作者合作关系。
关键词: 数据挖掘;链接分析 社团划分;可视化展示
For information in English, please click here
The Topological Characters and Community Structure of Co-authorship Networks
Abstract:In literature analysis area, the scale of data grows bigger and bigger, and data analysis based on tranditional methods becomes more and more difficult. At the same time, with the development of network science,scientists begin to consider the data processing technology based on linkage data. In this paper, we proposed a paragram which analyst a DBLP co-authorship. First, we gave out a list of basic statistics, such as degree distribution, weight of the links, coefficient, etc. And, we also analyzed the authors who participate in large amout of papers, tautonomy phenomenon, and frequent co-authorships. We found some interesting charictoristics through the analysis refered above. For example, the first several authors who have the most co-authors are all Chinese scientists with very short English names. The "person-centered network" of these scientists are all quit sparse, which implies significant tautonomy phenomenon. By comparing the two networks from Physics and biology, the network for Computer Science contains much lesser nodes. After delete the unstable relations which only appear once, we used some community detection method to the network, found out the relationship between different academic teams, and showd it visually. At the end, we discussed the internal structure and author relationship within a perticular team.
Keywords: Data Mining Link processing Community Detection Visualization
基金:
论文图表:
引用
No.4400778560402129****
同行评议
共计0人参与
勘误表
文献合作者网络中的拓扑特征与社团结构
评论
全部评论0/1000