Deep Learning-Based Model for Accurate Intrusion Detection in Network Traffic

Strategic Insights -

Researchers from the School of Computing Science and Engineering have developed an innovative deep learning-based model for intrusion detection in network traffic environments. This model combines a Convolutional Neural Network (CNN), Bidirectional Long Short-Term Memory (BiLSTM), and a self-attention mechanism to effectively capture spatial and temporal patterns while focusing on the most relevant features for intrusion detection. The proposed approach is rigorously evaluated on the NSL-KDD, UNSW-NB15, and IoTID20 datasets, demonstrating exceptional performance with high accuracy, precision, recall, and F1 score.

Key Takeaways:

  • The proposed model integrates a CNN, BiLSTM, and self-attention mechanism to capture spatial and temporal patterns in network traffic.
  • The model employs the Synthetic Minority Over-sampling Technique (SMOTE) with Edited Nearest Neighbors (ENN) to address the common issue of data imbalance and reduce bias toward majority classes.
  • Particle Swarm Optimization (PSO) is applied for hyperparameter tuning, enhancing the model's robustness and adaptability.
  • The model is rigorously evaluated on the NSL-KDD, UNSW-NB15, and IoTID20 datasets, demonstrating exceptional performance with high accuracy, precision, recall, and F1 score.
  • On the NSL-KDD dataset, the model achieved an accuracy of 99.93%, precision of 99.78%, recall of 99.60%, and F1 score of 99.65%.
  • On the UNSW-NB15 dataset, the model attained an accuracy of 99.70%, precision of 99.40%, recall of 99.50%, and F1 score of 99.35%.
  • On the IoTID20 dataset, the model achieved an accuracy of 99.78%, precision of 99.35%, recall of 99.38%, and F1 score of 99.40%.

Statistics:

  • Accuracy on the NSL-KDD dataset: 99.93%
  • Precision on the NSL-KDD dataset: 99.78%
  • Recall on the NSL-KDD dataset: 99.60%
  • F1 score on the NSL-KDD dataset: 99.65%
  • Accuracy on the UNSW-NB15 dataset: 99.70%
  • Precision on the UNSW-NB15 dataset: 99.40%
  • Recall on the UNSW-NB15 dataset: 99.50%
  • F1 score on the UNSW-NB15 dataset: 99.35%
  • Accuracy on the IoTID20 dataset: 99.78%
  • Precision on the IoTID20 dataset: 99.35%
  • Recall on the IoTID20 dataset: 99.38%
  • F1 score on the IoTID20 dataset: 99.40%

Sources:

  • Cluster Computing, 2025;28(11)
  • Springer, One New York Plaza, Suite 4600, New York, Ny, United States
  • School of Computing Science and Engineering, Cent South Univ, Changsha 410083, People's Republic of China
  • Irshad Ullah, Abida Naz, Kwizera K. Jonath, Muhammad Uzair, Abdul Haseeb Nizamani, and Husnain Mushtaq
  • NewsRx LLC
  • VerticalNews