Intrusion Detection System for IoT Based on Deep Learning and Modified Reptile Search Algorithm.

Area of Science:

• Cybersecurity
• Artificial Intelligence
• Network Security

Background:

• Intrusion detection systems (IDS) are crucial for securing Internet of Things (IoT) environments.
• Existing IDS often struggle with the high dimensionality and complexity of IoT data.
• Feature extraction and selection are vital for optimizing IDS performance.

Purpose of the Study:

• To propose a novel framework for enhancing IDS performance in IoT settings.
• To leverage deep learning and metaheuristic optimization for effective feature engineering.
• To improve the accuracy and efficiency of intrusion detection.

Main Methods:

• A framework combining Convolutional Neural Network (CNN) for feature extraction and Reptile Search Algorithm (RSA), a metaheuristic method, for feature selection.
• CNN learns relevant data representations in a reduced dimensional space.
• RSA identifies and selects the most impactful features from CNN outputs to optimize the IDS.

Main Results:

• The proposed framework demonstrated competitive classification performance across multiple benchmark datasets (KDDCup-99, NSL-KDD, CICIDS-2017, BoT-IoT).
• The RSA-based feature selection effectively identified optimal feature subsets, significantly boosting IDS performance.
• The framework achieved comparable or superior results to existing optimization methods in feature selection for IDS.

Conclusions:

• The novel framework integrating CNN and RSA offers a powerful approach to improve IDS performance in IoT environments.
• Metaheuristic optimization, specifically RSA, is highly effective for feature selection in complex network security data.
• This study provides a robust method for enhancing the detection capabilities of intrusion detection systems.