Enhancing IoT Network Security through Prompt Intrusion Detection Using Machine Learning

This study aims to enhance IoT network security by deploying rapid intrusion detection mechanisms fortified with machine learning techniques. Addressing the escalating security concerns surrounding IoT devices, the research develops effective strategies for swift intrusion identification and mitigation, encompassing various intrusion types such as Distributed Denial of Service (DDoS), Internet Control Message Protocol (ICMP), and Transmission Control Protocol Synchronize (TCP SYN). Leveraging supervised machine learning algorithms such as Support Vector Machines (SVM), Logistic Regression, Random Forest, and K-Nearest Neighbors (KNN), a highly accurate intrusion detection model is proposed. Evaluation of the model's performance, utilizing diverse datasets sourced from platforms like Kaggle, showcases notable accuracy rates across different intrusion types. Specifically, DDOS achieves 82% accuracy, TCP SYN attains 99.96%, and ICMP reaches 99.8% accuracy on average. Notably, Random Forest exhibits the highest accuracy among the tested algorithms. This research significantly contributes to strengthening IoT network security, bolstering overall resilience against malicious activities and unauthorized access.