Abstract: User-oriented online sharing communities have been development growth recently.The characteristics of these communities set them apart from traditional online message-based communities.This research constructs the user network using data collected from an IRC music channel from 2001 to 2006.After a series of measurements of the network，the authors represent the directed network as bipartite network.And then the authors detect the community structure by hierarchical clustering.According to the experiment result，the authors make an analysis of the shape，size and distribution of the community.Finally，the authors propose some approaches to enhance the net service.

Key words: complex network, community structure, clustering, social network

摘要： 用户为导向的在线分享型社区近几年发展迅速，而这些社区与以文字信息为基础的传统社区有较大的不同。利用IRC音乐频道从2001年到2006年的用户日志数据，构建了此共享型社区的用户网络。在对此网络的特点分析之后，将该网络的结构由有向的加权图转换为单向的二部图，从而利用层级聚类的方法，对网络中的用户群落进行了发掘。在此基础上，进一步分析了用户群落的大小，形状和分布的情况。最后根据研究的结果，提出了针对该社区提高服务的方法。

关键词: 复杂网络, 社区结构, 聚类, 社会网络