A Detection Method of Network Intrusion Based on SVM and Ant Colony Algorithm
Jianfeng Pu, Lizhi Xiao, Yanzhi Li, Xingwen Dong · 2012
This paper researches the intrusion detection problem of the network defense, pointing at the problem of low fitting defects in the traditional detection algorithm of high precision and low forecasting accuracy under the situation of small sample training, and puts forward the algorithm of Support Vector Machine.Aimed at the important influence of SVM kernel function on classification performance, this paper adopts the improved Ant Colony Algorithm as the method of selection SVM characteristics parameters.Experiments show this algorithm is significantly higher than the other algorithm in training and the detection speed, and have a high enhance of the detection rates of attacking sample.