Using Immune Algorithm to Optimize Anomaly Detection Based on SVM

In anomaly detection based on support vector machine, kernel parameter and error penalty c of support vector machine (SVM) determine generalization performance, and superfluous features of training samples affect classification performance. Thus, this paper presents a hybrid optimization selection method for SVM parameters and sample features using immune algorithm. Immune algorithms not only can convergence to global optimum, avoiding get in local optimum, but also can improve convergence rate. The experimental results show that our method can improve the classification accuracy and reduce the training time