Novel ML Approach for Securing Smart Home IoT Systems: Cyber Attack Detection Using ACI IoT Network Trafic Dataset −2023

Smart home systems, powered by Internet of Things (IoT) devices, are increasingly vulnerable to cyber-attacks due to their interconnected nature. This study proposed and implement an AI/ML based attack identification approach Integrated Probabilistic model for such systems by taking advantages of the robustness of multiple machine learning classifiers—Logistic Regression, k-Nearest Neighbors (KNN), Random Forest, Balanced Random Forest, Artificial Neural Networks (ANN), and Support Vector Machines (SVM). This research paper is using the ACI-IoT-2023 IoT network traffic dataset 2023. By simulating adversarial attacks on these binary classifiers, this research paper assesses their ability to detect and mitigate malicious activity in smart home environments. Our findings indicate varying levels of susceptibility across the models, with ensemble techniques like Balanced Random Forest and newly developed Integrated Probabilistic model showing greater resistance to adversarial inputs compared to other classifiers. The results of IPM integrate the results of different classifiers for improved security mechanisms to protect the smart homes network from increasingly sophisticated attacks with optimal false negative, accuracy, and f1 score of 0.98 and 0.99, respectively, and offering key insights for enhancing IoT system defenses.