Intrusion Detection System in IoT using Grey Wolf Optimization-Based Support Vector Machine

The Internet of Things (IoT) is appearing as a new technology for the development of different critical applications. To prevent adversarial attacks, fraud, and network intrusion, the Intrusion Detection System (IDS) has become a major component of organizations. In this research, the Grey Wolf optimization based Support Vector Machine (GWO-SVM) is proposed for the intrusion detection system using machine learning. Initially, the data is obtained by the Bot-IoT dataset and then min-max normalization is performed to normalize the acquired data. The different feature extraction approaches such as LeeNET, Gray-Level Co-occurrence matrix (GLCM), and Local Ternary Pattern (LTP) are used to extract appropriate features from the obtained data. The GWO approach is used for feature selection which examines appropriate features for classification. Finally, the SVM classification is performed to identify and classify intrusion detection accurately and effectively. The proposed GWO-SVM achieves a better accuracy of 99.67%, precision of 99.50%, recall of 99.47%, and f1-score of 99.60% respectively.