服务器通信网络中的业务系统发现方法研究

Based on the analysis we did on servers' communication network, we hope to find out the real individual business system consist of several servers from the massive and complicated communication network, by using method on complex network. Server is a physical carrier of applications, the business relationship between servers is directly reflected in the connections build by each of them. Because of the relations in one is relatively complex, but the relationship between the inside and outside is relatively sample. So finding out the business systems in a server communication network is a typical community detection problem. For this purpose, we presents a network modeling scheme, and the way to learn knowledge from the tag data. By comparing with the conventional methods of machine learning, proves that the weights learning model on relations of the servers we raised can improve the results of community detection algorithms.