Traffic Safety Management for Accident Casualties in Sudan: Characteristics, Evaluation and Prediction

Abstract

Recently there has been much concern about investigation of the problem of road traffic safety and, particularly prevention of traffic accidents. The importance of this research arose as a result of the excessive increase in traffic accidents and the resulting injuries and fatalities. Data from various sources were carefully examined and analyzed using ANN covering the period from 1991 to 2012, for comparative purposes Multiple Linear Regression (MLR) and Statistical Package for Social Science (SPSS) techniques were applied to fit the same data. Forecast for the years 2010 to 2020 were made using ANN, SPSS and MLR. The main objective of each of the methods was to fit a reliable prediction model. The adequacy of such model is typically measured either by the coefficient of determination (R2) as well as comparing actual accidents against predicted values. ANN resulted high values of R2 and better prediction capabilities.