Identification and ranking of accident black spots using advanced empirical Bayes method(PDF)
长安大学学报（自然科学版）[ISSN:1006-6977/CN:61-1281/TN]
- Issue:
- 2019年05期
- Page:
- 115-126
- Research Field:
- 交通工程
- Publishing date:
Info
- Title:
- Identification and ranking of accident black spots using advanced empirical Bayes method
- Author(s):
- TIAN Zhun; ZHANG Shengrui
- (1. School of Highway, Chang’an University, Xi’an 710064, Shaanxi, China; 2. School of Civil Engineering, Xi’an University of Architecture and Technology, Xi’an 710055, Shaanxi, China)
- Keywords:
- traffic engineering; traffic safety; signalized intersection; black spot identification and ranking; empirical Bayes; traffic accident statistical analysis
- PACS:
- -
- DOI:
- -
- Abstract:
- To accurately and effectively identify and rank traffic accident black spots, an accidentdatabased and an accidentpredictionmodelbased advanced empirical Bayes method was proposed. In addition, the practicability of the identification methods has been optimized. Furthermore, risk of accident involvement and safety benefits of treatment，two ranking criteria of black spots were suggested. Accident data for signalized intersections in two cities were collected. The accidentdatabased advanced empirical Bayes method was used to identify total accident black spots and fatal and injury accident black spots in city A. The total number of accident black spots identified in city A were subsequently ranked according to the risk of accident involvement. Concurrently, black spots in city B were identified by the accidentpredictionmodelbased advanced empirical Bayes method, and were ranked according to risk of accident involvement and safety benefits of treatment two criteria. The results show that 24 total accident black spots and 18 fatal and injury accident black spots were identified in city A at 95% confidence level, while 15 black spots were identified in city B at 99% confidence level. The advanced empirical Bayes method is superior to the rate quality control and accident rate methods regarding accuracy, the ability to eliminate the impact of random fluctuation and the regressiontothemean effect of accident frequency. The ranking criterion of the safety benefits of treatment favors the locations that are more costeffective to treat, while the ranking criterion of the risk of accident involvement favors the locations with higher deviation from the expected values. 7 tabs, 4 figs, 35 refs.
Last Update: 2019-10-16