Research intrusion detection based PSO-RBF classifier

In order to improve the accuracy of classification problem in intrusion detection, a hybrid classifier which was composed by KPCA, RBFNN and PSO, has been proposed in this paper. In the hybrid classifier, KPCA was used to reduce the dimensions, RBF was the core classification, and then PSO was used to optimize the parameters for EBFNN. The hybrid classifier used KPCA to extract the core nonlinear characteristics of raw data, introducing PSO to seek parameters overcame the weakness of RBFNN such as easily limit to local minimum points, low recognition rate and poor generalization. Finally the paper has done simulation using the KDDCUP99 data set in the matlab environment. Finally, the effectiveness of hybrid classifier was proved by experiments. Compared with traditional methods, the hybrid classifier has significantly improved the accuracy of classification in intrusion detection.