点击次数：

论文类型：期刊论文

发表时间：2010-11-01

发表刊物：Journal of Computational Information Systems

收录刊物：EI、Scopus

卷号：6

期号：11

页面范围：3567-3574

ISSN号：15539105

摘要：Recently, credit scoring has become a very important task and has been widely used in management science. In this study, we focused on the management of network, and studied a credit scoring model for network management basing on support vector machine (SVM). In our model, SVM is trained to classify the normal and abnormal network connection and using grid search algorithm to find optimum parameters for SVM model. We employ the decision distance of SVM model as the measurement of credit scoring. The advantage of our SVM based model is that the data can be classified effectively and less parameter need to be decided. With the network attack data set, the experimental result shows that our model can reduce the cost of network management and improve the performance of network management. ? 2010 Binary Information Press.