Optimizing Machine Learning Models for IoT-Based DDoS Attack Detection through Hyper parameter Tuning

Abstract

The proliferation of Internet of Things (IoT) devices has escalated the complexities and frequencies of Distributed Denial of Service (DDoS) attacks, making traditional detection mechanisms inadequate. This paper explores the enhancement of machine learning (ML) models specifically tuned for IoT environments using systematic hyperparameter optimization via grid search. By tailoring the learning processes and configurations of models such as Random Forest, XGBoost, and Support Vector Machines, the study achieves superior detection rates, reduced false positives, and improved computational efficiency. The findings suggest that precise hyperparameter tuning is crucial for adapting DDoS detection systems to the unique characteristics of IoT networks, thereby offering robust defenses against evolving cyber threats.