中国高校电子邮件网络实证研究

Empirical Analysis of an Email Network in a Chinese University

为了挖掘社会网络中人与人之间的交流方式,利用日志数据分析国内某高校电子邮件网络的特性。提取了两种网络:非交互网络和交互网络。结果发现两种网络均有很强的度和强度相关性,但是同配异配性质并不明显。此外,发现高强度节点之间存在富人俱乐部现象,而大度节点之间不存在此现象。最后利用加权适应度算法划分了加权网络的社团结构,发现社团规模累积分布满足幂律分布。结论为进一步建模仿真提供了实证基础。

To explore the communication pattern of individuals in social networks, by the log data we study the structural properties of an email network in a Chinese university. Two kinds of networks are constructed： unmutual network and mutual network. Both networks have a strong degree-strength correlation but neither of them has a significant assortative or disassortative property. Moreover, we find that there is a rich-club phenomenon among the high-strength nodes but not among the large-degree nodes. Finally, we detect the community structure of the weighted mutual network by weighted fitness algorithm and find that the cumulative distribution of community size follows power law. The results provide empirical foundation for model-building and simulations.