Employing Machine Learning for Predicting Transportation Modes under the COVID-19 Pandemic: A Mobility-Trends Analysis

ABSTRACT

With the advent of Coronavirus Disease 2019 (COVID-19), the world encountered an unprecedented health crisis due to the severe acute respiratory syndrome (SARS) pathogen. This impacted all of the sectors but more critically the transportation sector which required a strategy in the light of mobility trends using transportation modes and regions. We analyse a mobility prediction model for smart transportation by considering key indicators including data selection, processing and, integration of transportation modes, and data point normalisation in regional mobility. A Machine Learning (ML) driven classification has been performed to predict transportation modes efficiency and variations using driving, walking and transit. Additionally, regional mobility by considering Asia, Europe, Africa, Australasia, Middle-East, and America has also been analysed. In this regard, six ML algorithms have been applied for the precise assessment of transportation modes and regions. The initial experimental results demonstrate that the majority of the world's travelling dynamics have been contrastively shaped with the accuracy of 91.21% and 84.5% using Support Vector Machine (SVM) and Random Forest (RT) for different transportation modes and regions. This study will pave a new direction for the assessment of transportation modes affected by the pandemic to optimize economic benefits for smart transportation.