Network Intrusion Detection Using 1D Convolutional Neural Networks

The increasing complexity and volume of cyber threats necessitate advanced methods for ensuring computer network security. Traditional Network Intrusion Detection Systems (NIDS) often fall short in recognizing new and evolving threats. This research explores the use of 1D Convolutional Neural Network (1D-CNN), to enhance the effectiveness of NIDS. Utilizing the NF-UQ-NIDS-v2 dataset, the study demonstrates the potential of 1D-CNN models to accurately detect a wide range of network anomalies. The model achieves a high overall accuracy of 0.94, effectively identifying benign traffic, DDoS, DoS, scanning, and bot attacks. However, challenges remain in detecting infiltration, analysis, and worms. The findings emphasize the importance of integrating advanced AI technologies into NIDS to better respond to sophisticated cyber threats, particularly in regions like Indonesia, which recorded over 403 million network anomalies in 2023. This study provides valuable insights and a foundation for future advancements in network security.