An enhancing IOT security with a trinary deep learning paradigm and squirrel reptilian optimisation

This research work proposes a novel approach for IoT intrusion detection using a trinary deep learning paradigm. The proposed model aims to address the challenges of large data requirements and false positives commonly encountered in IoT network security. The model consists of four phases: pre-processing, multi-modal feature extraction, optimal feature selection, and intrusion detection. Initially, the collected raw data is pre-processed using data cleaning techniques and Z-score normalisation, which helps to standardise the data for further analysis. Following pre-processing, multi-modal feature extraction techniques are applied, including measures of central tendency, database features, statistical dispersion, and information entropy-based features. To select the most relevant features from the extracted set, the squirrel reptilian optimisation algorithm is employed. SRO combines the squirrel search algorithm and reptile search algorithm to optimise feature selection, ensuring that only the most informative features are utilised for intrusion detection.