Improving Network Security: An Intelligent IDS with RNN-LSTM and Grey Wolf Optimization

Abstract

While this dependence on interconnected computer networks and the web requires robust cybersecurity. Cyber threats have been met with solutions like Intrusion Detection Systems (IDS). IDS are commonly rule-based, and very often use either signature-based or heuristic approaches to detect intrusions. Therefore, for such we recommend an IDS that merges the Grey Wolf Optimization (GWO) algorithm and Recurrent Neural Networks with Long Short-Term Memory (RNN-LSTM). RNN-LSTM to Handle Dynamic Network data, but not provided enough complain details in model training. Based on the behavior of grey wolf, an optimization technique GWO is implemented for intrusion detection to enhance accuracy and minimize false alarm in RNN-LSTM. Preprocess and segment network data with creating RNN-LSTM model for considering the dependence of our dataset Our approach improves the IDS performance by optimizing hyperparameters such as hidden layers, units, learning rates using GWO. The architecture of this RNN-LSTM with GWO IDS provides capable and responsive intrusion detection, training on previous data to be able to detect new threats. Made for network security by combining deep learning and optimization, tests reached 99.5% accurate.

The research advances IDSs, addressing the limitations of traditional systems, and underscores the potential of AI and optimization in complex network security. This study demonstrates the promise of RNN-LSTM and GWO for creating robust, adaptive intrusion detection systems in intricate network environments.