Optimized Deep Learning Framework for Intrusion Detection in Network Traffic

With the improvement of the digital era comes an increased value for cybersecurity, and this needs to be addressed using advanced techniques. In this paper, we present an Intrusion Detection System (IDS) that combines the strengths of Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks, further optimized using Particle Swarm Optimization (PSO). The goal is to detect traffic patterns. The CNN components extract spatial features from the input data, while the LSTM captures the sequential dependencies, enhancing the detection of complex attack patterns. PSO is employed to automatically tune critical hyperparameters such as the number of LSTM units and dropout rate, improving both speed and classification accuracy. Experimental results demonstrate that the optimized model achieves an accuracy of 97.63%, outperforming traditional machine learning and non-optimized deep learning approaches. The proposed system provides a scalable and efficient solution for real-time intrusion detection in cyber environments.

@article{ijter2025212507255511,
  author = {Shaik Ishrath and I. Grace Asha Roy},
  title = {Optimized Deep Learning Framework for Intrusion  Detection in Network Traffic},
  journal = {International Journal of Technology &  Emerging Research },
  year = {2025},
  volume = {1},
  number = {3},
  pages = {161-168},
  issn = {3068-109X},
  url = {https://www.ijter.org/article/212507255511/optimized-deep-learning-framework-for-intrusion-detection-in-network-traffic},
  abstract = {With the improvement of the digital era comes an increased value for cybersecurity, and this needs to be addressed using advanced techniques. In this paper, we present an Intrusion Detection System (IDS) that combines the strengths of Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) networks, further optimized using Particle Swarm Optimization (PSO). The goal is to detect traffic patterns. The CNN components extract spatial features from the input data, while the LSTM captures the sequential dependencies, enhancing the detection of complex attack patterns. PSO is employed to automatically tune critical hyperparameters such as the number of LSTM units and dropout rate, improving both speed and classification accuracy. Experimental results demonstrate that the optimized model achieves an accuracy of 97.63%, outperforming traditional machine learning and non-optimized deep learning approaches. The proposed system provides a scalable and efficient solution for real-time intrusion detection in cyber environments. },
  keywords = {Intrusion Detection System, Convolutional Neural Networks, Long Short-Term Memory, Particle Swarm Optimization, Hyperparameters},
  month = {Jul},
}

Copyright © 2025 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/) which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.