Urban Road Accident Black Spot Identification and Classification Approach: A Novel Grey Verhuls–Empirical Bayesian Combination Method

Abstract

The identification and classification of accident black spots on urban roads is a key element of road safety research. To solve the problems caused by the randomness of accident occurrences and the unclear classification of accident black spots by the traditional model, we propose a method that can quickly identify and classify accident black spots on urban roads: a combined grey Verhuls–Empirical Bayesian method. The grey Verhuls model is used to obtain the predicted/expected numbers of accidents at accident hazard locations, and the empirical Bayesian approach is used to derive two accident black spot discriminators, a safety improvement space and a safety index (SI), and to classify the black spots into two, three, four and five levels according to the range of the SI. Finally, we validate this combined method on examples. High-quality and high-accuracy data are obtained from the accident collection records of the Ningbo Jiangbei District from March to December 2020, accounting for 90.55% of the actual police incidents during this period. The results show that the combined grey Verhuls–Empirical Bayesian method can identify accident black spots quickly and accurately due to the consideration of accident information from the same types of accident locations. The accident black point classification results show that the five-level rating of accident black points is most reasonable. Our study provides a new idea for accident black spot identification and a feasible method for accident black spot risk level classification.