群体是在线社交网络重要的中观组织。群体发现不仅有重要的理论意义,还推动了在线社交网络的应用与发展,有广泛的应用前景。该文总结论述了在线社交网络群体发现的研究进展。在分析群体形成机理的基础上定义在线社交网络群体,并介绍群体发现问题。根据挖掘群体时采用的不同特征,该文分别阐述基于个体属性特征的群体发现方法和综合属性与结构特征的群体发现方法。随后从特征选取和检测算法两个方面重点介绍了恶意行为群体的发现方法。最后,对群体发现进一步的研究方向进行展望。 Groups are important Meso-Organizations for online social networks. Group discovery not only has important theoretical significance, but also promotes the application and development of online social network, which has a wide range of application prospects. This article summarizes the research progress of the online social network group discovery. Define the online social network based on the analysis of the formation mechanism of the group and introduce the problem of group discovery. According to the different characteristics of mining groups, this paper elaborates the methods of group discovery based on individual attribute features and the methods of group discovery with integrated attributes and structural features. Followed by the feature selection and detection algorithms from two aspects focuses on the discovery of malicious behavior groups. Finally, the prospect of further research is found out.