Volume 8, Issue 4 (10-2012)
J Health Syst Res 2012, 8(4): 672-680 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Kazemi I, Tazhibi M, Momenyan S, Haghshenas H. تعیین عوامل مرتبط با تعداد تصادفات بر حسب شدت با استفاده از رگرسیون پواسون- لگ نرمال چند متغیره بیزی. J Health Syst Res 2012; 8 (4) :672-680
URL: http://hsr.mui.ac.ir/article-1-421-en.html
URL: http://hsr.mui.ac.ir/article-1-421-en.html
تعیین عوامل مرتبط با تعداد تصادفات بر حسب شدت با استفاده از رگرسیون پواسون- لگ نرمال چند متغیره بیزی
1- Assistant Professor, Department of Statistics, Department of Statistics, University of Isfahan, Isfahan, Iran
2- Assistant Professor, Department of Biostatistics and Epidemiology, School of Public Health, Isfahan University of Medical Sciences, Isfahan, Iran
3- MSc Student, Student Research Committee, Department of Biostatistics and Epidemiology, School of Public Health, University of Medical Sciences, Isfahan, Iran
4- Assistant Professor, Department of Civil Engineering, Department of Transportation, University of Isfahan, Isfahan, Iran
2- Assistant Professor, Department of Biostatistics and Epidemiology, School of Public Health, Isfahan University of Medical Sciences, Isfahan, Iran
3- MSc Student, Student Research Committee, Department of Biostatistics and Epidemiology, School of Public Health, University of Medical Sciences, Isfahan, Iran
4- Assistant Professor, Department of Civil Engineering, Department of Transportation, University of Isfahan, Isfahan, Iran
Abstract: (1252 Views)
Background: Univariate models have been used to analyse crash data in previous studies. With regards to the severity variable in these researches that can be more than one level, in the present study multivariate Poisson-log normal model has been recommended to model crash count by severity, although it has been also used to compare the univariate models. Methods: The statistical inference of model parameters was conducted by Bayesian method, via a Gibbs Sampler and the Metropolis-Hastings (M-H) algorithms. The crash data were related to the intersections of Isfahan, Iran. Findings: Despite over-dispersion at two levels of severity and correlation among these two levels, the multivariate Poisson-lognormal model offers a better fit than other models. Also the effect of the total AADT (average annual daily traffic) on the property was significant in all of the models. But the effect of total left turn AADT on injuries and fatalities was significant just in the univariate Poisson model. Under multivariate Poisson-lognormal model, an increase of 1000 vehicles in average total AADT was predicted to result in 31% more property damage. Conclusion: Reducing the total volume of traffic in the cost of accidents reduction was predicted to be very affordable in long term.
Type of Study: Research |
Subject:
education health and promotion
Received: 2020/07/16 | Accepted: 2012/10/15 | Published: 2012/10/15
Received: 2020/07/16 | Accepted: 2012/10/15 | Published: 2012/10/15
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
