A Multi-Stage Classification Approach for IoT Intrusion Detection Based on Clustering with Oversampling

Intrusion detection of IoT-based data is a hot topic and has received lot interests from researchers practitioners since the security IoT networks crucial. Both supervised unsupervised learning methods are used for intrusion networks. This paper proposes an approach three stages considering clustering with reduction stage, oversampling classification by Single Hidden Layer Feed-Forward Neural Network (SLFN) stage. The novelty resides in technique generating useful balanced training hybrid consideration detecting activities. experiments were evaluated terms accuracy, precision, recall, G-mean divided into four steps: measuring effect clustering, evaluation framework basic classifiers, technique, comparison classifiers. results show that SLFN choice Support Vector Machine Synthetic Minority Oversampling Technique (SVM-SMOTE) ratio 0.9 k value 3 k-means++ give better than other values techniques.