A Bayesian procedure for evaluating the frequency of calibration factor updates in highway safety manual (HSM) applications

详细信息    查看全文

• 作者：Dibakar Saha ; Ph.D.^a ; ^{dsaha003@fiu.edu} ; ^{debubuet111@gmail.com} (Research Associate) ; Priyanka Alluri ; Ph.D. ; P.E.^b ; ^{palluri@fiu.edu} (Assistant Professor) ; Albert Gan ; Ph.D.^a ; ^gana@fiu.edu (Professor)
• 关键词：Highway Safety Manual ; Calibration factor ; Bayesian estimation ; Marcov chain Monte Carlo ; Gibbs sampling ; Region of practical equivalence
• 刊名：Accident Analysis & Prevention
• 出版年：2017
• 出版时间：January 2017
• 年：2017
• 卷：98
• 期：Complete
• 页码：74-86
• 全文大小：897 K
• 卷排序：98

文摘

The variations between triennial, biennial, and annual calibration factors were assessed. A Bayesian estimation technique was applied to obtain comprehensive information about model parameters. The posterior distribution of the mean parameter was used to decide whether the variation is credibly high. Calibration factors need to be updated every two years, if variations not more than 0.01 are desired. Updating calibration factors at a frequency of three years is suggested for variations between 0.01 and 0.05.