Enhanced Network Intrusion Detection Using LSTM RNN

Abstract

For the security of networks from the external user or internal threats, Network intrusion detection systems (NIDS) are essential. Owing to the vast volume of network traffic, attacks are more likely to disrupt the network infrastructure or its users tremendously. Intrusion detection plays a crucial role in maintaining network stability by threats and harmful activities. Nonetheless, when designing effective IDSs for unexpected and abnormal assaults, there are several challenges. Deep learning techniques include a variety of tools, and known and unknown threats may be identified. Long Short-Term Memory (LSTM) is a type of RNN that can store values over arbitrary intervals. LSTM is a useful tool for classifying and detecting documented and unknown intrusions. In this review, we suggest a fundamental learning approach to IDS construction. We use LSTM RNNs and use NSL-KDD to train the pattern. Despite restricted computational power, the new model has achieved greater accuracy.