Smart Intrusion Detection using Supervised Learning and Feature Optimization

Abstract

In the digital age, safeguarding networks from cyber threats is critical. This project presents a network intrusion detection system (NIDS) using supervised machine learning (ML) techniques enhanced with feature selection strategies. By leveraging the NSL-KDD dataset, the study applies models like Support Vector Machine (SVM), Decision Tree, Logistic Regression, and Random Forest to detect malicious activities in network traffic. A correlation-based feature selection method optimizes the dataset by reducing redundancy, improving classification accuracy, and minimizing computational complexity. The performance is measured using standard evaluation metrics such as accuracy and F1-score. Among the tested algorithms, Random Forest achieved the highest detection rate, highlighting the system’s effectiveness in differentiating between normal and intrusive network behaviors.