Application on Network Traffic Prediction Based on Least Squares Support Vector Machine

Abstract

The network traffic is one of the important metrics for describing network behaviors, it plays an important role in network design, network protocol and traffic project implementation. In order to solve some problems in network traffic prediction, according to actual data for network- monitoring traffic, an approach to network traffic prediction is presented based on least squares support vector machine (LS-SVM), it mainly includes selecting for sample data of network traffic, normalization processing of data, network traffic model trained by LS-SVM and network traffic prediction, etc. Actual application results indicate that the method of network traffic prediction has high accuracy and good feasibility.