Intrusion Detection Model Based on Conditional Generative Adversarial Networks
Jingcheng Ye, Yunjie Fang, Junjie Ma · 2019
Aiming at the problem of low detection rate in the existing network intrusion detection model based on deep learning technology, and in order to avoid the phenomenon that can't converge or the convergence speed is too slow or even the model collapses due to the high degree of freedom of GAN, an intrusion detection model (IDM) based on CGAN is proposed. Different from the traditional generative adversary network (GAN), the model adds constraints to the discriminator and generator respectively to limit the degree of convergence freedom, so as to accelerate the effect of convergence. It almost completely retains the advantages of GAN and has a certain degree of optimization. The experimental results show that the network intrusion detection model based on CGAN method has obvious advantages in terms of accuracy, false alarm rate and real rate.