A copula-based approach for jointly modeling crash severity and number of vehicles involved in express bus crashes on expressways considering temporal stability of data.

The consequences of crashes, including injury, loss of lives, and damage to properties, are further worsened when buses plying expressways are involved in the crash. Previous studies have separately analyzed crash severity in terms of monetary cost, injuries and loss of lives, and the size of crashes in terms of the number of vehicles involved. However, as both outcome variables are correlated, it is imperative to perform a combined analysis using an appropriate econometric model to achieve a better model fit. This study contributes to the literature by jointly exploring the factors influencing the severity and size of express bus-involved crashes that occur on expressways and characterizes the dependence between both outcome variables by employing a more plausible copula regression framework. Likelihood ratio tests were also conducted to investigate the temporal stability of the factors that affect both crash severity and size. Based on the goodness-of-fit statistics, the Frank copula model proved superior to the independent ordered probit model. The estimate of the underlying dependence between the outcome variables provided a better comprehension of the correlation between them. Temporal instability was detected for the individual parameters in the models and is attributed to the changing driving behavior due to the heightened road safety campaigns. The results suggest that traffic exposure measures are significantly associated with a higher propensity of observing increased bus crash severity and size. Insights into the factors influencing the size and severity of express bus crashes are discussed, and appropriate engineering, enforcement, and education-related countermeasures are proposed.

Application of association rules mining algorithm for hazardous materials transportation crashes on expressway.

Although crashes involving hazardous material (HAZMAT) vehicles on expressways do not occur frequently compared with other types of vehicles, the number of lives lost and social damage is very high when a HAZMAT vehicle-involved crash occurs. Therefore, it is essential to identify the leading causes of crashes involving HAZMAT vehicles and make specific countermeasures to improve the safety of expressways. This study aims to employ the association rules mining (ARM) approach to discover the contributory crash-risk factors of HAZMAT vehicle-involved crashes on expressways. A case study is conducted using crash data obtained from the Korea Expressway Corporation crash database from 2008 to 2017. ARM was conducted using the Apriori algorithm, and a total of 855 interesting rules were generated. With appropriate support, confidence, and lift values, we found hidden patterns in the HAZMAT crash characteristics. The results indicate that HAZMAT vehicle-involved crashes are highly associated with male drivers, single vehicle-involved crashes, clear weather conditions, daytime, and mainline segments. Also, we found that HAZMAT tank-lorry and cargo truck crashes, single vehicle-involved crashes, and crashes on mainline segments of expressways had independent and unique association rules. The finding from this study demonstrates that ARM is a plausible data mining technique that can be employed to draw relationships between HAZMAT vehicle-involved crashes and significant crash-risk factors, and has the potential of providing more easy-to-understand results and relevant insights for the safety improvement of expressways.

Endogenous commercial driver's traffic violations and freight truck-involved crashes on mainlines of expressway.

Freight truck-involved crashes result in a high mortality rate and significantly impact logistic costs; therefore, many researchers have analyzed the causes of truck-involved traffic crashes. In the existing literature, it was found that truck-involved crashes are affected by factors such as road geometry, weather, driver and vehicle characteristics, and traffic volume based on a variety of statistical methodologies; however, the endogenous impact resulting from driver traffic violation has not been considered. The goal of the study is to discover the factors influencing freight vehicle crashes and develop more accurate crash probability estimation by explaining the endogenous driver traffic violations. To achieve the purpose of this study, we applied the two-stage residual inclusion (2SRI) approach, a methodology used in the nonlinear regression analysis model for capturing the endogeneity issue. This method improves the accuracy of the model by capturing the unobserved effects of driver traffic violations. From the results, traffic violations were identified to be influenced by the driver's physical condition, as well as driver and vehicle characteristics. Furthermore, variables of driver traffic violations such as improper passing, speeding, and safe distance violation were found to be endogenous in the probability model of freight truck crashes on expressway mainlines.