Improving intrusion detection system accuracy using deep neural network

Abstract

Internet of Things (IoT) has emerged as an intelligent network that connects objects to the Internet, allowing them to interact with each other without human intervention. The accessibility of IoT devices through unprotected networks subjected them to security vulnerabilities and various malicious attacks. While traditional Intrusion Detection Systems were introduced to address IoT security issues, there is need for intelligent intrusion detection methods. This study attempts to address and mitigate these security challenges by enhancing the performance and efficiency of IDSs with proposed Deep Neural Network (DNN) model. The study use a Deep Neural Network (DNN) and processed IoTID20 datasets for the detection of intrusion. The performance of the system is evaluated using performance metrics; Accuracy, Precision, recall and F1-Score. The optimal result accuracy obtained is 99.04%. The proposed model has demonstrated a potential improvement of Intrusion Detection Systems.