Bagging Ensemble Model Performance in IoT Cyberattack Detection: A Comprehensive Evaluation

Fast growth of Internet of Things (IoT) devices has presented novel challenges in ensuring network security. With advancements in technology, cyberattacks such as Mirai, Distributed Denial of Service, and spoofing have become more sophisticated, posing significant risks to IoT systems. These attacks disrupt network operations and IoT services, creating serious obstacles to the reliable functioning and adoption of IoT devices. The expansion of IoT devices has raised the potential of cyberattacks, requiring efficient and dependable detection techniques. This study focuses on the use of machine learning techniques for the detection of IoT issues, analyzing the efficacy of different machine learning models like Logistic Regression, Random Forest, Gradient Boosting, AdaBoost, and a Bagging Ensemble (BE) model on UNSWNB15 dataset in which selectKbest method is utilized for feature selection and Smote technique to handle imbalance dataset. The models were assessed using various measures, including accuracy, precision, recall, F1 score, and false negative rate (FNR). However, the BE model outperformed all others, achieving an accuracy of 97.18%, and a low FNR of 3.70%. The findings underscore the efficacy of bagging ensemble method, in enhancing classification performance by relying on the advantages of individual models. This paper highlights the performance of bagging ensemble method as effective solutions for IoT attack detection, providing high accuracy and diminished error rates.