Machine Learning-based Services Provisioning for Intelligent Internet of Vehicles

Abstract

This paper is aimed to deliver a Machine Learning (ML) based intelligent system that is capable of intelligently issuing services in a pre-defined environment setup that simulates a simple real-life scenario of Internet of Vehicle (IoV). First, a detailed discussion about Vehicular Ad Hoc Networks (VANETs) and IoVs is introduced stating the significant differences between both of them and why IoVs outplay VANETs. A thorough literature review about the fundamental aspects of IoV is clearly addressed. Following the literature review, an environment setup is constructed backed up with an empirically generated dataset. This then paves the way to examine two different Machine Learning classifiers, namely Binary Logistic Regression and Shallow Neural Network for our ML based intelligent system. Both classifiers are discussed in terms of mechanism and mathematical formulation. Finally, an analysis of both classifiers’ performance along with the necessary statistical measures are presented and discussed in addition to a conclusive comparison between both classifiers.