Intelligent Detection of Network Agent Behavior Based on Support Vector Machine

摘要

Aiming at the illegal agent behaviors in current network, a new intelligent recognition method based on support vector machine(SVM) is proposed for network agent behavior. This method selects RBF(Radial Basic Function) as the kernel function for SVM classifier, and applies the SVM active learning algorithm to the detection of network agent behavior. Through the effective learning of SVM, ordinary data and network agent behavior data can be distinguished correctly. Then an intelligent detection mechanism is established, which takes SVM as the active learning machine. The mechanism can detects network access behavior and identifies network agent behavior effectively. In this way, the source of network agent behavior can be located accurately and timely, and the monitoring of network traffic can be complished finally.