earticle

영어

It is sought to develop a proper model which considers a dispersion feature of count data with discrete probability distribution, and simultaneously overcomes the shortcoming of the linear regression model, widely adopted for analyzing traffic accidents. By utilizing the survey data on company taxi drivers in Jeonju, Iksan and Kunsan cities, overdispersion and heterogeneity existing in those data are tested by statistical verification and then the optimal model with higher reliance and accuracy is proposed. Also, Sensitivity analysis, called as marginal effects, is performed to know the change of traffic accidents depending on the change in drivers' socio-economic characteristic and other elements. It should be stressed to analyze primary factors of taxi accidents whose rate is higher than any other vehicle in order to diminish the entire traffic accidents in the cities as well as to gain trust in public transportation. This study is expected to help building fundamental countermeasures to prevent or reduce traffic accidents by analyzing casual relationship of behavioral and environmental aspects influencing taxi accidents.