Intrusion Detection Method Based On Improved Conditional Generative Adversarial Network

Abstract

At present, the intrusion detection model mainly uses anomalous behavior to establish a library of intrusion behavior patterns, and determines whether the intrusion behavior conforms to the intrusion behavior specification by comparing the library of intrusion behavior patterns. Once there is a change in intrusion behavior or a new type of network attack, the existing intrusion detection model cannot make corresponding changes according to the actual changes. Therefore, making intrusion detection models have the ability to learn autonomously and be able to adapt to changes in the network environment to detect new types of unknown attacks has received increasing attention from many security researchers. In this paper, we propose an intrusion detection model (CGAN-RF) based on conditional generative adversarial network (CGAN) and random forest (RF). The CGAN-RF model improves the class imbalance problem of the dataset by generating samples to enhance the detection efficiency of minority and unknown classes.