LightEnsemble-Guard: An Optimized Ensemble Learning Framework for Securing Resource-Constrained IoT Systems

Light Ensemble-Guard is an optimized ensemble learning framework designed to enhance the security of resource-constrained Internet of Things (IoT) systems. With the rapid growth of IoT devices and their integration into critical infrastructure, ensuring their security has become increasingly challenging due to limited computational resources. This framework leverages a combination of lightweight machine learning models in an ensemble structure to deliver accurate and efficient threat detection while minimizing the computational footprint. By incorporating multiple base models, including decision trees, random forests, and XGBoost classifiers, LightEnsemble-Guard offers superior detection accuracy for various IoT security threats, such as DoS (Denial of Service) attacks, malware, and unauthorized access. The framework uses an optimized stacking method to enhance prediction accuracy and reduce model complexity, making it suitable for deployment in resource-constrained environments. Furthermore, it incorporates advanced data preprocessing techniques to handle noisy IoT data and ensures real-time detection capabilities. The evaluation of LightEnsemble-Guard across multiple attack scenarios demonstrates its robustness and efficiency, providing a scalable solution for IoT security. This research aims to empower IoT systems with advanced security features without compromising performance, ensuring safe and reliable operation of IoT devices in critical sectors.

Keywords: IoT Security, Ensemble Learning, Machine Learning, XGBoost, Random Forest, Decision Tree, Stacking Classifier, Threat Detection, Resource-Constrained Systems, DoS Attacks, Lightweight Models, Cybersecurity