Convolutional Neural Network for Advanced Intrusion Detection for Data Balancing on System Efficiency

In 2024, the telecommunications sector experienced a significant surge in cyber threats, with BT detecting approximately 2,000 potential cyber-attack signals per second, marking a 1,200% increase in malicious scanning activity compared to the previous year. Concurrently, the healthcare industry faced a 32% rise in weekly attacks, averaging 2,018 per organization from January to September 2024. These escalating threats underscore the critical need for robust Network Intrusion Detection Systems (NIDS). However, the effectiveness of machine learning-based NIDS is significantly hindered by the inherent class imbalance in network intrusion datasets, leading to poor detection of minority class attacks. The proposed system CNNs for efficient feature extraction and attack categorization, while a hybrid data balancing approach, combining under-sampling of majority classes and over-sampling of minority classes, is employed to address dataset imbalance. The methodology is validated using the NSL-KDD and BoT-IoT datasets, with performance evaluated in terms of recall, False Alarm Rate (FAR), and F1-score for both binary and multi-class classification. Experimental results demonstrate that the proposed system significantly improves minority attack detection while maintaining a low false positive rate, outperforming conventional oversampling techniques such as SMOTE, ADASYN, and Generative Adversarial Networks (GANs). The findings underscore the effectiveness of hybrid data balancing in enhancing deep learning-based NIDS, thereby improving overall threat detection and cybersecurity resilience.