The alarming increase in the number of road traffic accidents, in particular casualty or injury accidents, in Mauritius affirms our understanding of road traffic injuries as a major national health problem. In order to bring down these numbers effectively, it is pertinent to scientifically model the effects of the factors leading to traffic accidents on the roads of Mauritius. The PF 178 form consists of important information on each accident across Mauritius. However, these information have not been adequately utilised so far to research on the major causes of road accidents. In this study, we first convert these information into a proper data structure and use Generalized Linear Models (GLMs) and Artificial Neural Network (ANN) approaches that can efficiently identify the significant factors underlying road traffic accidents and predict the severity of these accidents. The inferential results illustrate that the types of road structures, the day and time effect, street lighting conditions, vehicle types and conditions and driver profiles are the potential influential factors in the causation of road traffic accidents from 2012-2017. The findings of this research would, thus, be of utmost benefit to the concerned authorities and policymakers. In this way, this would formulate and enforce the appropriate preventive measures while simultaneously strengthening the current traffic system.
Keywords
Road traffic accidents,GLM and ANN approaches,Estimation of effects,Predictions.