An Improved LSTM Network Intrusion Detection Method
Liang Zhang, Hao Yan, Qingyi Zhu
Abstract
The characteristics of high network traffic dimension and large data volume make the traditional network intrusion detection model have a longer response time, lower detection accuracy, and seriously endanger the data security of network entities. In order to solve this problem, this paper studies the improved LSTM intrusion detection algorithm model, and uses Quantum Particle Swarm Optimization (QPSO) to select the network traffic data to reduce the feature dimension. The dimensionality-reduced network traffic is classified to detect network intrusion behavior. After testing on the KDDCup99 data set, the experimental results show that the QPSO feature selection algorithm can select the optimal feature subset, and the improved LSTM network can effectively improve the accuracy and F1-Score of intrusion detection.