A Robust Intrusion Detection In Network Traffic Using Deep Learning

Intrusion detection is a core topic in network security, and intrusion detection algorithms based on deep learning have become a research hotspot in network activity.In our proposal, we construct a network intrusion detection classification model (CNN-LSTM) based on a deep learning model to sense anomalies in network activity and effectively distinguish and identify network traffic data to ensure the security of clients' private data.The CNN-LSTM model combines the powerful learning ability of long short-term memory neural networks with time series data.It learns and classifies features selected by convolutional neural networks, and verifies applicability based on multi-classification scenarios.The model is trained using classic datasets NSL-KDD and UNSW_NB15.The verification accuracy and training loss on both datasets show good convergence and stability.Additionally, the accuracy rate is high when classifying various types of traffic.Experimental results show that the overall performance of the model has been significantly improved, demonstrating a high detection rate, classification accuracy, and low false alarm rate