Enhanced Intrusion Detection in IoT with a Novel PRBF Kernel and Cloud Integration

Abstract

The proliferation of Internet of Things (IoT) devices in various sectors has increased the need for robust security solutions capable of addressing complex network vulnerabilities and sophisticated cyber threats. This study introduces a novel architecture that integrates cloud computing with advanced machine learning techniques to provide efficient and scalable security in IoT systems. A unique Polynomial Radial Basis Function (PRBF) kernel is proposed to enhance the classification accuracy of Support Vector Machine (SVM) beyond traditional Gaussian and polynomial kernels. This study compares the proposed PRBF-SVM with Logistic Regression, SVM, and XGBoost models, optimized through rigorous hyperparameter tuning, to demonstrate significant improvements in detection rates. Furthermore, the integration of cloud services facilitates the offloading of computationally intensive tasks, ensuring scalability and real-time response capabilities. The results highlight the superior performance of the proposed model in accuracy, efficiency, and computation time, making a compelling case for its application in safeguarding IoT environments against evolving threats.