基于自组织特征映射神经网络的校园网用户聚类分析

An Analysis of Campus Network Users' Behaviors Based on Self Organization Map

在校园网系统的管理中,网管人员常常需要了解用户的网络行为,以便更好地配置和管理网络系统,提高网络服务效率。以大理学院校园网认证计费系统日志为研究对象,利用自组织映射神经网络(SOM)对校园网用户行为和网络通信模式进行聚类分析。网管人员可以根据聚类结果,设计有效的网络管理方案和系统配置参数,从而有效地配置和管理校园网,提高服务效率和管理质量。

The network administrators often want to understand the users＇ network behaviors in order to further control their operation and behavior effectively. The paper proposes a method for the analysis of users＇ network behaviors based on Self Organization Map （SOM）. The proposed method utilizes the SOM to analyze users＇ network behavior and traffic patterns based on authentication and charge log data of IP networks. The results of the analysis can be used as assistance for the administrators to design efficient controlling and configuring parameters of network management systems.