Deep learning-based early stage detection (DL-ESD) for routing attacks in Internet of Things networks

Security represents one of the main critical issues in the Internet of Things (IoT), especially the routing attacks in the core network where the loss of information becomes very harmful. This paper proposes a novel scheme called deep learning-based early stage detection (DL-ESD) using IoT routing attack dataset (IRAD), including hello flood (HF), decreased rank (DR), and version number (VN) to enhance the detection capability of routing attacks. The experiments have been performed in three phases: (i) features extraction using linear discriminant analysis (LDA), which aims to generate features more distinguishable from each other, (ii) the features normalization using min-max scaling to eliminate the worst overfittings to the existence of fewer data points in training samples, and (iii) selection the substantial features. The binary classification methods have been employed to measure the proposed model's training efficiency. We have performed the training stage on deep learning techniques such as logistic regression (LR), K-nearest neighbors (KNN), support vector machine (SVM), naive Bayes (NB), and multilayer perceptron (MLP). The comparison results illustrate that the proposed MLP classifier has a high training accuracy and the best runtime rate. Consequently, the proposed scheme achieved prediction accuracy reaching 98.85%, precision of 97.50%, recall rate 98.33%, and 97.01% F1 score rate with better performance than state-of-the-art studies.