Investigating the Cumulative Distribution of Accident Severity on Chinese Roads

Road traffic accidents remain a major concern in China due to an astonished number of casualties. To better investigate the cumulative regularity of accidents at different severity levels, a novel method considering cumulative probability of severity distribution is introduced. By using Heinrich’s theory, a 7-year accident dataset collected in China is applied to analyze the relationship between accidents and number of fatalities. Given the positive-skewed characteristic of accident data, the inverse functions of the cumulative distribution function of the chi-square distribution, F distribution and Weibull distribution are used to establish the statistical model to describe the distribution regularity of road traffic accidents. The results reveal that when road traffic accidents in China are classified by fatalities, the accident data conforms to the characteristics of positive-skewed distribution. The best degrees of freedom of the chi-square distribution and F distribution are 0.16, 19, and 12, respectively, and the scale parameter and shape parameter of Weibull distribution are 0.7 and 1.1, which are obtained by fitting the accident data. The test results of goodness-of-fit by minimum mean square error, F(19,12) is more reasonable to explain the distribution of accident severity on Chinese roads from 2005 to 2011, and the fitting error within an acceptable range.