AN ANALYTICAL STUDY OF HYBRID MACHINE LEARNING TECHNIQUES FOR INTRUSION DETECTION AND PREVENTION IN THE INTERNET OF THINGS

Abstract

This paper states an analytic study of Hybrid machine learning (ML) and deep learning (DL) techniques that aims to solve the growing security challenges of the Internet of Things (IoT) ecosystem. The rapid proliferation of connected devices and the low latency requirements of IoT systems mean that traditional security mechanisms often struggle to balance high detection accuracy and low latency. We propose a new Four-Level Hybrid Security Framework consisting of a combination of anomaly-based and signature-based detection along with a multi-phased risk factor analysis. By processing a meta-analysis of recent performance information for 2024-2025, we are able to show that the proposed hybrid approach is significantly more efficient in terms of detection rates for zero-day attacks and computational efficiency. Our findings produce a strong blueprint for developing new-age Intrusion Detection and Prevention Systems (IDPS) that are scalable and resilient to evolving cyber threats.