ENSEMBLE TECHNIQUE BASED MODEL FOR INTRUSION DETECTION IN INTERNET OF THINGS ENVIRONMENT

Abstract

ABSTRACT

The rapid growth of interconnected devices gives rise to possible exploits and compromise of devices in Internet of Things (IoTs) environment, which will require an established fact for such compromise to be checkmate. This research focuses on detection of threat attack detection in IoT environment based on ensemble technique. In the proposed ensemble model, an outlier analysis was performed in order to optimize the features for enhanced model performance, while Support Vector Machine and Feed Forward Neural Network serves as the base learners which were combined to form the proposed ensemble model, this was employed in order to enhance performance strength of training model in detection of intrusion threat in IoT environment. The advantage of the proposed model is its ability to generate an enhanced performance evaluation output; as the result proves an excellent performance for intrusion detection in IoTs environment. The obtained accuracy, precision, F-score, and recall of the proposed ensemble model are 99.96 for each respectively outperforming the existing technique with a record of 92.42 and 95.43 for accuracy and detection rate respectively, this is a clear distinction of the superiority of ensemble (Feed Forward Neural Network and Support Vector Machine) model over traditional model, of which the proposed ensemble model can serve as technique in forestall intrusion threat experienced in IoTs environment.