DEEP LEARNING MODEL FOR INTRUSION DETECTION SYSTEM IN IOT NETWORK

Abstract

 The Internet of Things (IoT) has revolutionized modern life by enabling seamless communication between devices and servers. However, the rapid growth of IoT has made it a target for cyber threats, necessitating the development of effective Intrusion Detection Systems (IDS) to ensure network security. Deep learning, with its advanced capabilities, has emerged as a powerful tool in improving IDS performance. This paper proposes an advanced deep learning model tailored for detecting anomalies in IoT networks. The framework incorporates sparse autoencoders for feature extraction and utilizes a Support Vector Machine (SVM) classifier for accurate intrusion detection. Validation of the model is conducted using the KDD-DSN dataset. To gauge its effectiveness, metrics such as accuracy, precision, recall, F-score, ROC-AUC. The results are rigorously compared to existing intrusion detection methods, highlighting the proposed framework's superior accuracy and efficiency in identifying threats within IoT environments.