An Intrusion Detection Method based on Stacked Autoencoder and Support Vector Machine

Abstract

To explore the feasibility of forming a combined classifier for intrusion detection by applying Support Vector Machine (SVM) and Stacked Autoencoder (SAE), an intrusion detection method based on Stacked Autoencoder and Support Vector Machine is proposed. Considering that the Piecewise Radial Basis Function (P-RBF) in the Support Vector Machine can improve the classification performance, the Radial Basis Function (RBF) and P-RBF are selected for the proposed method, and the detection performance of the above two methods are compared. Experiments show that the method based on Stack Autoencoder and Support Vector Machine using P-RBF kernel are superior to the one using RBF kernel in accuracy, detection rate and false alarm rate. The method based on Stack Autoencoder and Support Vector Machine provides an idea for the application in the field of intrusion detection.