International Journal of Highway Engineering (한국도로학회논문집)

Korean Society of Road Engineers (한국도로학회)
권호 표지
DOI QR코드

DOI QR Code

Safety Performance Functions for Central Business Districts Using a Zero-Inflated Model

영과잉을 고려한 중심상업지구 교통사고모형 개발에 관한 연구

  • 이상혁 (한국건설기술연구원 도로연구소) ;
  • 우용한 (경일대학교 건설공학부)
  • Received : 2016.05.24
  • Accepted : 2016.07.15
  • Published : 2016.08.16
Download PDF
⟨ Previous Next ⟩

Abstract

PURPOSES : The purpose of this study was to develop safety performance functions (SPFs) that use zero-inflated negative binomial regression models for urban intersections in central business districts (CBDs), and to compare the statistical significance of developed models against that of regular negative binomial regression models. METHODS : To develop and analyze the SPFs of intersections in CBDs, data acquisition was conducted for dependent and independent variables in areas of study. We analyzed the SPFs using zero-inflated negative binomial regression model as well as regular negative binomial regression model. We then compared the results by analyzing the statistical significance of the models. RESULTS : SPFs were estimated for all accidents and injury accidents at intersections in CBDs in terms of variables such as AADT, Number of Lanes at Major Roads, Median Barriers, Right Turn with an Exclusive Turn Lane, Turning Guideline, and Front Signal. We also estimated the log-likelihood at convergence and the likelihood ratio of SPFs for comparing the zero-inflated model with the regular model. In he SPFs, estimated log-likelihood at convergence and the likelihood ratio of the zero-inflated model were at -836.736, 0.193 and -836.415, 0.195. Also estimated the log-likelihood at convergence and likelihood ratio of the regular model were at -843.547, 0.187 and -842.631, 0.189, respectively. These figures demonstrate that zero-inflated negative binomial regression models can better explain traffic accidents at intersections in CBDs. CONCLUSIONS : SPFs that use a zero-inflated negative binomial regression model demonstrate better statistical significance compared with those that use a regular negative binomial regression model.

Keywords

References

  1. Carlin, B. P., and Louis, T. A., 2009. Bayesian Methods for Data Analysis, New York, NY, CRC Press.
  2. Chin, H. C., and Quddus M. A., 2003. Applying the Random Effect Negative Binomial Model to Examine Traffic Accident Occurrence at Signalized Intersection, Accident Analysis and Prevention, vol. 35, No. 2. 253-259. https://doi.org/10.1016/S0001-4575(02)00003-9
  3. Jonsson, T., Lyon, J., Washington, S., Schlakwyk, I., and Lord, D., 2009. Differences in the Performance of Safety Performance Functions Estimated for Total Crash Count and for Crash Count by Crash Type, Transportation Research Record, 2102, 115-123.
  4. Jovanie, P. P., and Chang, H., 1986. Modeling the Relationship of Accidents to Miles Traveled, Transportation Research Record, 1068. 42-51.
  5. Korean National Police Agency, 2015. Traffic Accident Statistics.
  6. Kweon, Y., 2012. Disaggregate Safety Evaluation for Signalized Intersections and Evaluation Tool, Journal of Advanced Transportation, Vol. 46, No. 2. 151-160. https://doi.org/10.1002/atr.150
  7. Lee, S. H., Lee, Y. D., and Do, M. S., 2012. Safety Impacts of Red Light Enforcement on Signalized Intersections, Journal of Korean Society of Transportation, Vol. 30, No. 6. 93-102.
  8. Lee, S. H., Lee, Y. D., and Do, M. S., 2016. Analysis on Safety Impact of Red Light Cameras using the Empirical Bayesian Approach, Transportation Letters, Online Published.
  9. Lee, S. H., Park, M. H., and Woo, Y. H., 2015. A Study on Developing Crash Prediction Model for Urban Intersections Considering Random Effects, Journal of the Korea Institute of Intelligent Transport Systems, Vol. 14, No. 1. 85-93. https://doi.org/10.12815/kits.2015.14.1.085
  10. Lord, D., Washington, S. P., and Ivan, J. N., 2005. Poisson, Poisson-Gamma and Zero-Inflated Regression Models of Motor Vehicles: Balancing Statistical Fit and Theory, Accident Analysis and Prevention, Vol. 37. 35-46. https://doi.org/10.1016/j.aap.2004.02.004
  11. Miaou, S. P., and Lum, H. A., 1993. Modeling Vehicle Accidents and Highway Geometric Design Relationship, Accident Analysis and Prevention, Vol. 25, No. 6. 689-709. https://doi.org/10.1016/0001-4575(93)90034-T
  12. Mitra, S., and Washington, S., 2012. On the Significance of Omitted Variables in Intersection Crash Modeling, Accident Analysis and Prevention, Vol. 49, 439-448. https://doi.org/10.1016/j.aap.2012.03.014
  13. Mouatassim, Y., and Ezzahid, E. H., 2012. Poisson Regression and Zero-inflated Poisson Regression: Application to Private Health Insurance Data, European Actuarial Journal, Vol. 2, No. 2, 187-204. https://doi.org/10.1007/s13385-012-0056-2
  14. Na, H., and Park, B. H., 2012. Accident Models of Circular Intersection by Cause Using ZAM, International Journal of Highway Engineering, Vol. 14, No. 2. 101-108.
  15. Retting, R. A., Ulmer, R. G., and Williams, A. F., 1999. Prevalence and Characteristics of Red Light Running Crashes in the United States, Accident Analysis and Prevention, Vol. 31. 687-694. https://doi.org/10.1016/S0001-4575(99)00029-9
  16. Vuong, Q. H., 1989. Likelihood Ration Tests for Model Selection and Non-nested Hypotheses, Econometrica, Vol. 57, No. 2, 307-333. https://doi.org/10.2307/1912557

Cited by

  1. A Study on the Improvement of Prediction Accuracy for Traffic Accident Models Using Machine Learning (Generalized Regression Neural Network) vol.20, pp.6, 2018, https://doi.org/10.7855/IJHE.2018.20.6.179