Accident Prediction Models for Winter Road Safety

Abstract

Most accident prediction models are developed with single-level count data models, such as the traditional negative binomial models with fixed or varying dispersion parameters, assuming independence of data. For many accident data sets in road safety analysis, especially those that are highly disaggregated (hourly data), a hierarchical structure in the data often manifests in some form of correlation. Crash prediction models developed with aggregate data could produce biased results because of the assumption of data independence and inflation of the adequacy of the model's explanation because of the use of aggregate data. The potential effects of data aggregation and correlation on accident prediction models are investigated. The analysis uses an accident database that includes hour-level and storm-level accident counts for individual winter snowstorms at four highway sections in Ontario, Canada. Models of two levels of aggregation, aggregated event-based models and disaggregated hourly based models, were developed. The effect of data aggregation had a significant effect on model results, whereas the difference between conventional regression and multilevel regression was inconsequential.

Developing Commercial Motor Vehicle Crash-Specific Safety Performance Functions at Interchange Ramp Terminals in Kentucky

“…A prominent extension of the NB model is the HTNB model, which is a more flexible approach to address unobserved heterogeneity through allowing dispersion parameter to vary across crash sites as a function of site-specific characteristics. The superiority of HTNB model over the traditional NB model has been confirmed in previous studies ( 12 , 13 , 24 , 25 ). Similar to the traditional NB model, the HTNB model uses the same probability density function as that given in Equation 2.…”

Bayesian analysis of factors affecting crash frequency and severity during winter seasons in Iowa

“…Note that these results are similar with that of a past study Hermans et al (2006a) using data from 37 sites. However, our results are different from the study conducted by Usman (2011) that found a negative relationship between visibility and crash frequency during a storm event. Note that the models developed in this research are not winter-storm event based models but rather consider all the weather-related crashes that occurred during the winter seasons.…”

“…The results make intuitive sense as higher wind speeds could cause blowing snow effects, which might impair driver performance during winter seasons. This result is in agreement with previous research findings ( Usman, 2011).…”

“…Using disaggregate hourly data from the same winter snow storm events as in Usman et al (2010) , Usman et al (2011) developed a generalized negative binomial model for predicting winter crash frequency. This model was compared to the model calibrated from using aggregate event-based data to examine the impact of data aggregation (from event based data to hourly data) on modeling results.…”

Improvements of Warning Signs for Black Ice Based on Driving Simulator Experiments

“…The occurrence time and location of this natural phenomenon are extremely difficult to predict because rapid transitions between dry, wet, snow, and ice are common on such surfaces. This recurrent condition caused by water, ice, and snow causes many accidents worldwide [ 1 , 2 , 3 , 4 , 5 ]. Given the terrain in Korea, many roadways pass through mountainous areas [ 6 ].…”