期刊
ANALYTIC METHODS IN ACCIDENT RESEARCH
卷 24, 期 -, 页码 -出版社
ELSEVIER
DOI: 10.1016/j.amar.2019.100110
关键词
Driver injury severity; Temporal indicator; Unobserved heterogeneity; Latent class; Marginal effect
资金
- National Natural Science Foundation of China [61808083]
- Postdoctoral Science Foundation of China [2018M630497]
Temporal variation has been recognized as one of the major sources of unobserved heterogeneity in traffic safety research that has not been completely addressed. Overlooking temporal variation may result to biased estimates of effects of impact factors. This paper develops a latent class mixed logit model with temporal indicators to investigate highway single-vehicle crashes and the effects of significant contributing factors to driver injury severity. Crash data from 2010 to 2016 in Washington State is collected with a total of 31,115 single-vehicle crashes. The developed model is able to interpret both within- and across-class unobserved heterogeneity and temporal variation. After a careful comparison, a two-class model is selected as the final model. Estimation results show that: two temporal indicators show significant influence on latent class probability functions; urban indicator and principle type are found to be random parameters and have significant heterogeneity in the mean as a function of male indicator and driver's age indicators. Variables with fixed effects, including animal collision, overturn collision, off-road collision, winter, minor arterial, interstate, wet, snow, ice, speed limit, vehicle age, turning movement, out control movement, lane-change movement, no airbag, deployed airbag, partial and total ejection, seatbelt, and no liability, show significant impacts on different levels of injury severity outcomes in each class. This study provided an insightful understanding of the time-varying effects of the significant factors on driver injury severity using marginal effect analysis, and the temporal indicators in the proposed model were found to enhance the model capability of temporal variation identification. (C) 2019 Elsevier Ltd. All rights reserved.
作者
我是这篇论文的作者
点击您的名字以认领此论文并将其添加到您的个人资料中。
推荐
暂无数据