Artificial Neural Networks and Support Vector Machine for IoT

Abstract

AbstractThe continuous expansion and advances of low-power wireless communication practices enable the manifestation of numerous novel appliances in IoT. Wireless tools have a higher hand than wired technology and made a global impact on approximately each business segment in terms of cost, robustness, and traceability. The future Internet of Things enhances the progression of various advanced information services that are cost-effective, accessible from anywhere, and ubiquitous. However, on the other hand, expanding IoT, there are countless security and protection challenges because of improved attack pieces of knowledge and increased number of related resources. To provide well-organized, mostly constant, little latency transportation and persistent connectivity for the Internet-of-Things (IoT) devices, machine learning (ML) and deep learning (DL) models were exploited.First, we present a detailed view of ML and DL applicability in WSN and IoT. Then we converse a complete view of various neural networks (NN) and support vector machine (SVM) types that incorporate frequent, deep neural networks, quarter and ellipsoidal SVMs, and subspace-SVM forms, which are relevant to wireless and IoT appliances. Then, we provide an in-depth summary of various communication issues in IoT that are addressed by neural networks and SVM and application and motivation for using those techniques. Followed by intrusion detection in IoT with NN and SVM, a case study on outlier detection WSNs data and future research implementations is discussed.KeywordsInternet of ThingsSVMANNEnergy harvestingDeep learningRNNCNN