A Novel Intrusion Detector Based on Deep Learning Hybrid Methods
Shizhao Wang, Chunhe Xia, Tianbo Wang · 2019
Intrusion detection system plays an important role in network security defense. It analyzes network traffic and connection characteristics to identify various types of network attacks. Deep learning based intrusion detectors perform better in predicting unknow attacks and detection accuracy. In this paper, we use long short-time memory (LSTM) in recurrent neural network (RNN) units, and propose an improved long short-time memory tree (LSTMTree) model with ability of secondary detection to solve the problem of high false negative rate in the RNN intrusion detector. And then we use NSL-KDD data set to verify the performance of our presented model. The experimental results show that our model can improve the detection performance better than previous models.