AnAlgorithm of Intrusion Detection Based on RS and WSVM
Yuan-Zheng Cheng · Jisuanji fangzhen · 2011
Algorithms of network intrusion detection are discussed.To solve the problems of current machine learning methods,such as decreased generalization capacities under incomplete training samples,low speed with large scale samples and high system resource consumption,a new method based on rough set and wavelet kernel SVM is proposed.Firstly,the attributes and the number of training samples are reduced by rough set to decrease samples' scale.Then the wavelet kernel LS-SVM is trained by the samples and the classification model is build.At last,the test samples are detected.Results of the simulation show that contrast to traditional methods,this method achieves higher detection speed,lower false rate and resource consumption.