Supervised Learning Approach for Intrusion Detection System in Imbalanced Network

Abstract

In the era of increasing cyber threats, Intrusion Detection Systems (IDS) play a crucial role in safeguarding network infrastructures. Traditional IDS solutions often struggle to detect sophisticated attacks, particularly in the presence of highly imbalanced network traffic where malicious activities are significantly outnumbered by normal behavior. This study presents a supervised learning-based approach for intrusion detection that addresses the challenges posed by imbalanced datasets. Benchmark datasets such as NSL-KDD and CICIDS are utilized for model training and evaluation. Various supervised learning algorithms, including Decision Trees, Random Forests, Support Vector Machines, and Artificial Neural Networks, are employed to classify network traffic. To tackle the imbalance issue, data-level techniques like SMOTE (Synthetic Minority Over-sampling Technique) and algorithm-level approaches such as cost-sensitive learning are incorporated. Performance is evaluated using precision, recall, F1-score, and ROC-AUC to ensure effectiveness in identifying minority class intrusions. Results demonstrate that combining supervised learning with imbalance handling significantly improves detection accuracy for rare attack types. This work highlights the importance of model selection and data preprocessing in enhancing IDS performance. Future directions include leveraging deep learning and real-time traffic analysis to further strengthen intrusion detection capabilities in complex and dynamic network environments.