A Machine Learning and Deep Learning Approach to Network Intrusion Detection System

Network Intrusion Detection Systems (NIDS) have been a vital tool for protecting computer networks from security threats, such as unauthorized access and malicious attacks through monitoring and analyzing network activities, and alerting administrators to unusual behavior. In this project, we explored both Machine Learning (ML) and Deep Learning (DL) techniques to improve the effectiveness of NIDS systems. Our approach includes Random Forest (RF), Support Vector Machine (SVM), Extreme Gradient Boosting (XGBoost), and a Hybrid LSTM (H-LSTM) model that incorporates Convolutional Neural Networks (CNN). Our study shows that among these techniques, XGBoost delivers a high accuracy (99.80\%) in intrusion detection, demonstrating it as an important technique for improving the efficiency and reliability of NIDS systems.