Server Load Prediction Based on Improved Support Vector Machines

摘要：

To provide e-learning service more efficiently and effectively, Data mining technique have been applied in web-based distance education such as personalized service provision, server load prediction, etc.In web-based e-learning system, web server is the key and core component.In this paper,a novel server load prediction model is put forward by employing Support Vector Machines (SVM). In addition, an approach to select free parameters of SVM is introduced which select parameters by checking if the training residual is white noise. Theoretical analysis and Experimental result has shown that by using this approach, server load prediction with high precision can be achieved.

关键词： server load Support Vector Machines whitenoise

作者: Yanhua Yu Xiaosu Zhan Junde song

作者单位: The school of Electronic Engineering,Beijing University of Posts and Telecommunications,Beijing 1000876,China

会议类型: 国际会议

会议名称: 2008 IEEE International Symposium on IT in Medicine and Education(2008信息技术在医学和教育中的应用国际研讨会)(ITME 2008)

会议地点: 厦门

会议语种:英文

页码: 838-842

在线出版日期: 2008-12-12（万方平台首次上网日期，不代表论文的发表时间）