Ensemble based Intrusion Detection System for IoT Device
P. Ananthi, T.E. Ramya, R. Janani · 2023
An Intrusion Detection System (IDS) is a security-based mechanism used for finding malicious activities and unauthorized access in a computer network or system. IDS solutions are now more essential than ever because of the rising use of Internet of Things (IoT) devices. This study proposes an approach to build IDS for IoT devices is to use Recursive Feature Elimination (REE) with KDD 99 dataset. REE is a feature selection algorithm that removes irrelevant or redundant features recursively until the optimal feature subset are obtained. The KDD 99 dataset contains network traffic data that a proposed learning model can be trained on to classify normal and malicious network traffic. A deep neural network is deployed for classification after the necessary features have been selected via RER A deep neural network can be trained on the selected features to categorise network traffic as legitimate or malicious using the KDD99 dataset. The proposed IDS performance is assessed by using various evaluation metrics such as recall, precision, F1-score and accuracy, The Hyper parameter tuning and ensemble learning are used to improve the performance representations of IDS.