RFSVMDD: Ensemble of multi-dimension random forest and custom-made support vector machine for detecting RPL DDoS attacks in an IoT-based WSN environment

Abstract

Growing dependence on the internet of things (IoT) and wireless sensor networks (WSNs) has led to critical security issues, especially concerning distributed denial of service (DDoS) attacks based on RPL. Such attacks can severely compromise the network’s security, reliability, and efficiency. To effectively address this problem, this research proposes (RFSVMDD) a novel hybrid detection model that combines a multi-dimensional random forest (MDRF) with a custom-made support vector machine (CSVM). The proposed technique uses MDRF to provide scalability for consistent feature selection and anomaly detection across high-dimensional datasets. CSVM reduces false positives and increases detection accuracy through its improved classification of potential threats. Experimental assessments in simulated IoT-based WSN environments show that the model outperforms conventional machine learning methods regarding accuracy, detection speed, and durability. This novel ensemble approach presents a promising solution by enhancing IoT and WSN networks against RPL DDoS attacks.