EACE Journal — Vol. 12, No. 2
Mathematical Modeling of the Correlation between Frequency of Road Traffic Accidents and Contributing Factors
Tarekegn Shirko Lachore
Abstract
The study investigated the causes and contributing factors of a road traffic accident. In order to achieve the stated objective, the total number of 532 road crash records for a period of six years (2011 - 2016) was collected from secondary sources (traffic police records and transport office database) throughout the entire road. A questionnaire survey, field measurement, and observations were also carried out to collect the required data for accident analyses. Based on statistical analysis, the significant variables with strong prediction ability for model development were selected among three categorized contributing factors. The mathematical models were developed using partial least square regression and multiple nonlinear regression techniques. In order to determine the interactive effect of individual variables on the model, the Monte Carlo approach was employed with the 10,000 times repeated simulation of samples within the probability density functions of the input data. The results revealed that the education level of driver and driving experience has a high contribution to traffic accident in relation to the human factors, with an absolute value of 55.47% and 18.01%. In the road geometric factors, the gradient, sight distance, and horizontal curve were highly correlated with accident frequency and have a correlation coefficient of 0.692, -0.529, and -0.426 respectively. In addition, vehicular factors like the service period of a vehicle, type of vehicle, and volume has a significant effect on the road traffic accident
Read Full Article (PDF)