Intrusion Detection Model Based on Improved Support Vector Machine

Abstract

With development and popularization of computer network, network security problems increasingly bring into prominence. Intrusion detection technique can effectively enlarge the scope of protection on network and system. An intrusion detection method based on support vector machine (SVM) is studied. Aiming at the shortcoming of SVM on detecting precision, an intrusion detection model based on improved SVM is put forward according to hypothesis test theory. To confirm the effectiveness of this approach, a simulation testing is done. The experiment results show that the improved SVM has stronger learning ability and higher accuracy and lower false positive rate.