Analytic Methods in Accident Research最新文献

英文中文

A note on data segmentation, sample size, and model specification for crash injury severity modeling 关于碰撞损伤严重程度建模的数据分割、样本量和模型规范的说明

IF 12.5 1区工程技术 Q1 PUBLIC, ENVIRONMENTAL & OCCUPATIONAL HEALTH

Analytic Methods in Accident Research

Pub Date : 2025-02-12 DOI: 10.1016/j.amar.2025.100373

Qinzhong Hou , Jinglun Zhuang , Chenrui Zhai , Xiaoyan Huo , Fred Mannering

In recent years, the statistical assessment of crash injury severity data has increasingly begun to segment the available crash data into observational groups to explore the possibility that such groups may share the same estimated parameters. This method is commonly used to account for parameters that may shift over time, where the data is often segmented into groups based on observational year. Unfortunately, such data segmentation can lead to small samples within each group, which has caused some concern about decreasing sample size. However, concerns about diminishing sample size are often misplaced and not well understood. In this paper, the impact of data segmentation is assessed by estimating models that address the possibility of temporally shifting parameters. Starting with a large 80,000 observation sample, the process involves randomly segmenting the data into groups with sample sizes varying from 1000 to 40,000, and then assessing the difference between the estimated data-segmented models and the overall model (using all available data) using likelihood ratio tests. The results indicate that: 1) model specification is extremely important, regardless of sample size, 2) statistical tests should be used to determine the suitability of simple versus complex models, not sample size, and 3) the variance/covariance structure of the data being considered determines model specification and sample size effects, which means sample-size requirements are data-specific, and that general statements regarding minimum sample size requirements for specific model types cannot be made.

近年来，碰撞损伤严重程度数据的统计评估越来越多地开始将可获得的碰撞数据划分为观察组，以探索这些组可能具有相同估计参数的可能性。这种方法通常用于解释可能随时间变化的参数，其中数据通常根据观测年份分成几组。不幸的是，这样的数据分割可能导致每个组内的小样本，这引起了一些关于减少样本量的担忧。然而，对样本量减少的担忧往往是错误的，也没有得到很好的理解。在本文中，通过估计模型来评估数据分割的影响，该模型解决了暂时转移参数的可能性。从80,000个大型观察样本开始，该过程涉及将数据随机分割为样本量从1000到40,000不等的组，然后使用似然比检验评估估计的数据分割模型与总体模型（使用所有可用数据）之间的差异。结果表明：1)模型规格极其重要，无论样本量如何；2)统计检验应用于确定简单模型与复杂模型的适宜性，而不是样本量；3)所考虑的数据的方差/协方差结构决定模型规格和样本量效应，这意味着样本量要求是针对具体数据的，不能就具体模型类型的最小样本量要求作出一般性说明。

{"title":"A note on data segmentation, sample size, and model specification for crash injury severity modeling","authors":"Qinzhong Hou , Jinglun Zhuang , Chenrui Zhai , Xiaoyan Huo , Fred Mannering","doi":"10.1016/j.amar.2025.100373","DOIUrl":"10.1016/j.amar.2025.100373","url":null,"abstract":"<div><div>In recent years, the statistical assessment of crash injury severity data has increasingly begun to segment the available crash data into observational groups to explore the possibility that such groups may share the same estimated parameters. This method is commonly used to account for parameters that may shift over time, where the data is often segmented into groups based on observational year. Unfortunately, such data segmentation can lead to small samples within each group, which has caused some concern about decreasing sample size. However, concerns about diminishing sample size are often misplaced and not well understood. In this paper, the impact of data segmentation is assessed by estimating models that address the possibility of temporally shifting parameters. Starting with a large 80,000 observation sample, the process involves randomly segmenting the data into groups with sample sizes varying from 1000 to 40,000, and then assessing the difference between the estimated data-segmented models and the overall model (using all available data) using likelihood ratio tests. The results indicate that: 1) model specification is extremely important, regardless of sample size, 2) statistical tests should be used to determine the suitability of simple versus complex models, not sample size, and 3) the variance/covariance structure of the data being considered determines model specification and sample size effects, which means sample-size requirements are data-specific, and that general statements regarding minimum sample size requirements for specific model types cannot be made.</div></div>","PeriodicalId":47520,"journal":{"name":"Analytic Methods in Accident Research","volume":"45 ","pages":"Article 100373"},"PeriodicalIF":12.5,"publicationDate":"2025-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143430171","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

How do drivers manage speed at tunnel entrances? Insights from uncorrelated grouped random parameters duration models for model invalidation and performance recovery times 司机如何管理隧道入口的车速？从不相关的分组随机参数持续时间模型中了解模型失效和性能恢复时间

IF 12.5 1区工程技术 Q1 PUBLIC, ENVIRONMENTAL & OCCUPATIONAL HEALTH

Analytic Methods in Accident Research

Pub Date : 2025-01-23 DOI: 10.1016/j.amar.2025.100371

Yunjie Ju , Shi Ye , Tiantian Chen , Guanyang Xing , Feng Chen

Human drivers must quickly adjust to perturbations at tunnel entrances (i.e., the rapid switching of cross-sections, abrupt longitudinal changes in the driving environment, and changes in visual illumination, denoted “tunnel transition perturbations”) to regain control of their vehicles, especially when managing speed to prevent motor overshoot. Previous research has assessed drivers’ visual adaptation rather than variations in vehicle control under tunnel transition perturbations. In this study, a sample entropy method was used to measure the safety–critical duration of speed control events at tunnel entrances and thereby reveal the participants’ speed adaptation and recovery performance under tunnel transition perturbations. Two key metrics—model invalidation time and performance recovery time—were introduced, and an uncorrelated grouped random parameters hazard-based duration model was developed. Road grade, road curvature, income, and time having held a license were positively associated with model invalidation time, while a history of accidents in the past 12 months was negatively associated with model invalidation time. In addition, road grade, road curvature, and income had heterogeneous effects on model invalidation time. Moreover, a history of accidents in the past 12 months moderated the relationship between road grade and model invalidation time. Furthermore, road curvature, average annual mileage, and sleep deprivation significantly influenced performance recovery time, while road grade and non-fatigue condition had heterogeneous effects on performance recovery time. Overall, this study demonstrated that the participants’ personal characteristics and experiences significantly shaped the development of their internal models, and that their current status and perception had a substantial influence on their performance recovery under tunnel transition perturbations. These insights enhance understanding of the mechanisms of drivers’ motor control under tunnel transition perturbations and will therefore enable improvement of road traffic design and safety management at tunnel entrances.

人类驾驶员必须迅速适应隧道入口的扰动（即，横断面的快速切换，驾驶环境的突然纵向变化以及视觉照明的变化，称为“隧道过渡扰动”），以重新控制车辆，特别是在管理速度以防止电机超调时。先前的研究评估的是驾驶员的视觉适应，而不是隧道过渡扰动下车辆控制的变化。本研究采用样本熵法测量隧道入口速度控制事件的安全临界持续时间，从而揭示隧道过渡扰动下参与者的速度适应和恢复性能。引入了两个关键指标——模型失效时间和性能恢复时间，并建立了一个不相关的分组随机参数基于风险的持续时间模型。道路等级、道路曲率、收入和持有驾照的时间与车型失效时间呈正相关，而过去12个月内的交通事故历史与车型失效时间呈负相关。此外，道路坡度、道路曲率和收入对模型失效时间的影响存在异质性。此外，过去12个月的事故历史缓和了道路等级与模型失效时间之间的关系。此外，道路曲率、平均年里程和睡眠剥夺对性能恢复时间有显著影响，而道路等级和非疲劳状态对性能恢复时间有异质性影响。总体而言，本研究表明，参与者的个人特征和经历显著地塑造了他们的内部模型的发展，他们的现状和感知对他们在隧道转换扰动下的绩效恢复有实质性的影响。这些见解增强了对隧道过渡扰动下驾驶员运动控制机制的理解，因此将有助于改进隧道入口的道路交通设计和安全管理。

{"title":"How do drivers manage speed at tunnel entrances? Insights from uncorrelated grouped random parameters duration models for model invalidation and performance recovery times","authors":"Yunjie Ju , Shi Ye , Tiantian Chen , Guanyang Xing , Feng Chen","doi":"10.1016/j.amar.2025.100371","DOIUrl":"10.1016/j.amar.2025.100371","url":null,"abstract":"<div><div>Human drivers must quickly adjust to perturbations at tunnel entrances (i.e., the rapid switching of cross-sections, abrupt longitudinal changes in the driving environment, and changes in visual illumination, denoted “tunnel transition perturbations”) to regain control of their vehicles, especially when managing speed to prevent motor overshoot. Previous research has assessed drivers’ visual adaptation rather than variations in vehicle control under tunnel transition perturbations. In this study, a sample entropy method was used to measure the safety–critical duration of speed control events at tunnel entrances and thereby reveal the participants’ speed adaptation and recovery performance under tunnel transition perturbations. Two key metrics—model invalidation time and performance recovery time—were introduced, and an uncorrelated grouped random parameters hazard-based duration model was developed. Road grade, road curvature, income, and time having held a license were positively associated with model invalidation time, while a history of accidents in the past 12 months was negatively associated with model invalidation time. In addition, road grade, road curvature, and income had heterogeneous effects on model invalidation time. Moreover, a history of accidents in the past 12 months moderated the relationship between road grade and model invalidation time. Furthermore, road curvature, average annual mileage, and sleep deprivation significantly influenced performance recovery time, while road grade and non-fatigue condition had heterogeneous effects on performance recovery time. Overall, this study demonstrated that the participants’ personal characteristics and experiences significantly shaped the development of their internal models, and that their current status and perception had a substantial influence on their performance recovery under tunnel transition perturbations. These insights enhance understanding of the mechanisms of drivers’ motor control under tunnel transition perturbations and will therefore enable improvement of road traffic design and safety management at tunnel entrances.</div></div>","PeriodicalId":47520,"journal":{"name":"Analytic Methods in Accident Research","volume":"45 ","pages":"Article 100371"},"PeriodicalIF":12.5,"publicationDate":"2025-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143103716","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Understanding the effects of underreporting on injury severity estimation of single-vehicle motorcycle crashes: A hybrid approach incorporating majority class oversampling and random parameters with heterogeneity-in-means 了解漏报对单辆摩托车碰撞伤害严重程度估计的影响：一种结合多数类过抽样和随机参数的混合方法

IF 12.5 1区工程技术 Q1 PUBLIC, ENVIRONMENTAL & OCCUPATIONAL HEALTH

Analytic Methods in Accident Research

Pub Date : 2025-01-23 DOI: 10.1016/j.amar.2025.100372

Nawaf Alnawmasi , Apostolos Ziakopoulos , Athanasios Theofilatos , Yasir Ali

The underreporting of crash data is a well-documented issue in road safety literature, but few studies have focused on addressing this problem in the context of analyzing crash injury severities. This paper aims to provide an empirical assessment of the impact of underreporting issue using a hybrid approach in estimating injury severity for single-vehicle motorcycle crashes. Unlike traditional machine learning methods that oversample the minority class (the category with the fewer observations such as fatal and severe injuries), the present study oversamples the majority class (i.e. minor injuries), which are often underreported in crash datasets, thus providing a fresh perspective on this issue. Afterwards, random parameter models with heterogeneity in means and variances were applied. The results of this study, as supported by the likelihood ratio tests, indicate that the key variables influencing motorcyclists’ injury severities remain consistent across both original and oversampled data models. Specifically, crashes occurring during slowing down or stopping are associated with lower injury severity, whereas negotiating a right turn increases the probability of severe injuries. Interestingly, crashes that occur on dry pavements are associated with higher injury severity when compared to wet pavements, likely due to rider behavior adjustments in adverse weather conditions to compensate for the risk. Overall, the oversampled models have a significantly lower marginal effects values compared to the original model’s marginal effects. This study provides a foundation for further examination of underreporting issue in crash injury severity modelling and also highlights the need to capture the dynamics of crash injuries suggesting that alternative approaches could improve the understanding and hence road safety management. Future studies are encouraged to replicate this methodology to validate the findings as well as utilize other advanced machine learning algorithms, like tree-based models to assess underreporting mitigation.

在道路安全文献中，碰撞数据的漏报是一个有充分记录的问题，但很少有研究集中在分析碰撞伤害严重程度的背景下解决这个问题。本文旨在使用混合方法对漏报问题的影响进行实证评估，以估计单车摩托车碰撞的伤害严重程度。与传统的机器学习方法(对少数类（致命和严重伤害等观察较少的类别）进行过采样不同，本研究对大多数类（即轻伤）进行过采样，这在碰撞数据集中经常被低估，从而为这个问题提供了一个新的视角。然后，采用均值和方差均异质性的随机参数模型。本研究的结果得到似然比检验的支持，表明影响摩托车手伤害严重程度的关键变量在原始和过抽样数据模型中保持一致。具体来说，在减速或停车时发生的撞车事故与较低的受伤严重程度有关，而右转则增加了严重受伤的可能性。有趣的是，与湿路面相比，在干燥路面上发生的撞车事故与更高的伤害严重程度有关，这可能是由于骑手在恶劣天气条件下调整行为以补偿风险。总体而言，过采样模型的边际效应值明显低于原始模型的边际效应值。这项研究为进一步研究碰撞伤害严重程度建模中的漏报问题提供了基础，也强调了捕捉碰撞伤害动态的必要性，这表明替代方法可以提高对碰撞伤害的理解，从而提高道路安全管理。鼓励未来的研究复制这种方法来验证研究结果，并利用其他先进的机器学习算法，如基于树的模型来评估低报缓解。

{"title":"Understanding the effects of underreporting on injury severity estimation of single-vehicle motorcycle crashes: A hybrid approach incorporating majority class oversampling and random parameters with heterogeneity-in-means","authors":"Nawaf Alnawmasi , Apostolos Ziakopoulos , Athanasios Theofilatos , Yasir Ali","doi":"10.1016/j.amar.2025.100372","DOIUrl":"10.1016/j.amar.2025.100372","url":null,"abstract":"<div><div>The underreporting of crash data is a well-documented issue in road safety literature, but few studies have focused on addressing this problem in the context of analyzing crash injury severities. This paper aims to provide an empirical assessment of the impact of underreporting issue using a hybrid approach in estimating injury severity for single-vehicle motorcycle crashes. Unlike traditional machine learning methods that oversample the minority class (the category with the fewer observations such as fatal and severe injuries), the present study oversamples the majority class (i.e. minor injuries), which are often underreported in crash datasets, thus providing a fresh perspective on this issue. Afterwards, random parameter models with heterogeneity in means and variances were applied. The results of this study, as supported by the likelihood ratio tests, indicate that the key variables influencing motorcyclists’ injury severities remain consistent across both original and oversampled data models. Specifically, crashes occurring during slowing down or stopping are associated with lower injury severity, whereas negotiating a right turn increases the probability of severe injuries. Interestingly, crashes that occur on dry pavements are associated with higher injury severity when compared to wet pavements, likely due to rider behavior adjustments in adverse weather conditions to compensate for the risk. Overall, the oversampled models have a significantly lower marginal effects values compared to the original model’s marginal effects. This study provides a foundation for further examination of underreporting issue in crash injury severity modelling and also highlights the need to capture the dynamics of crash injuries suggesting that alternative approaches could improve the understanding and hence road safety management. Future studies are encouraged to replicate this methodology to validate the findings as well as utilize other advanced machine learning algorithms, like tree-based models to assess underreporting mitigation.</div></div>","PeriodicalId":47520,"journal":{"name":"Analytic Methods in Accident Research","volume":"45 ","pages":"Article 100372"},"PeriodicalIF":12.5,"publicationDate":"2025-01-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143103711","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Time-dependent effect of advanced driver assistance systems on driver behavior based on connected vehicle data 基于车联网数据的高级驾驶辅助系统对驾驶员行为的时变影响

IF 12.5 1区工程技术 Q1 PUBLIC, ENVIRONMENTAL & OCCUPATIONAL HEALTH

Analytic Methods in Accident Research

Pub Date : 2025-01-15 DOI: 10.1016/j.amar.2025.100370

Yuzhi Chen , Yuanchang Xie , Chen Wang , Liguo Yang , Nan Zheng , Lan Wu

This paper proposes a novel functional data analysis approach to investigate the time-dependent effect of advanced driver assistance systems (ADAS), specifically forward collision warnings, on driver speed reduction behavior. Existing aggregate measures compress temporal information within driver behavior profiles and fail to explicitly reveal the temporal dependency of such effect. With the proposed approach, the functional representation method is adopted to capture the underlying driver behavior in response to warning messages and address issues of irregularly spaced observations and measurement errors; the results of the functional principal component analysis with the bootstrap-enhanced Kaiser-Guttman method reveal important patterns in driver response behaviors; and a nonparametric functional varying coefficient regression model, considering vehicle initial motions and drivers’ acceleration styles, is established. This regression model utilizes coefficient functions to estimate the time-dependent effect of ADAS. The proposed approach is evaluated based on the New York City connected vehicle dataset using forward collision warning event records. The results suggest that the treatment effect of the warning messages is time-dependent, initially increasing before progressively decreasing over time. Driver responses can be decomposed into several phases at the 95 % confidence level, including reaction time (1.3 s), brake adjustment time (1.3 s), progressive braking duration (2.7 s), and effective treatment duration (4.0 s). The time-dependent bootstrap confidence interval confirms driver heterogeneity in these distinct phases. The proposed functional data analysis approach can serve as a paradigm for quantifying the treatment effect of other ADAS applications. The findings can support the improvements of ADAS design and the development and calibration of driver behavior models accounting for ADAS.

本文提出了一种新的功能数据分析方法来研究先进驾驶辅助系统（ADAS）对驾驶员减速行为的时间依赖性影响，特别是前向碰撞警告。现有的聚合度量压缩了驾驶员行为概况中的时间信息，并且不能明确地揭示这种影响的时间依赖性。该方法采用函数表示方法捕捉驾驶员响应警告信息的潜在行为，并解决观测间隔不规则和测量误差的问题；基于自举增强的Kaiser-Guttman方法的功能主成分分析结果揭示了驾驶员响应行为的重要模式；建立了考虑车辆初始运动和驾驶员加速方式的非参数变系数函数回归模型。该回归模型利用系数函数来估计ADAS的时变效应。基于纽约市互联汽车数据集，使用前向碰撞预警事件记录对该方法进行了评估。结果表明，警告信息的治疗效果是时间依赖性的，最初增加，然后随着时间的推移逐渐减少。在95%的置信水平下，驾驶员的反应可以分解为几个阶段，包括反应时间（1.3 s）、制动调整时间（1.3 s）、渐进制动时间（2.7 s）和有效处理时间（4.0 s）。随时间变化的自举置信区间证实了驾驶员在这些不同阶段的异质性。所提出的功能数据分析方法可以作为量化其他ADAS应用的治疗效果的范例。研究结果可为ADAS设计的改进以及考虑ADAS的驾驶员行为模型的开发和校准提供支持。

{"title":"Time-dependent effect of advanced driver assistance systems on driver behavior based on connected vehicle data","authors":"Yuzhi Chen , Yuanchang Xie , Chen Wang , Liguo Yang , Nan Zheng , Lan Wu","doi":"10.1016/j.amar.2025.100370","DOIUrl":"10.1016/j.amar.2025.100370","url":null,"abstract":"<div><div>This paper proposes a novel functional data analysis approach to investigate the time-dependent effect of advanced driver assistance systems (ADAS), specifically forward collision warnings, on driver speed reduction behavior. Existing aggregate measures compress temporal information within driver behavior profiles and fail to explicitly reveal the temporal dependency of such effect. With the proposed approach, the functional representation method is adopted to capture the underlying driver behavior in response to warning messages and address issues of irregularly spaced observations and measurement errors; the results of the functional principal component analysis with the bootstrap-enhanced Kaiser-Guttman method reveal important patterns in driver response behaviors; and a nonparametric functional varying coefficient regression model, considering vehicle initial motions and drivers’ acceleration styles, is established. This regression model utilizes coefficient functions to estimate the time-dependent effect of ADAS. The proposed approach is evaluated based on the New York City connected vehicle dataset using forward collision warning event records. The results suggest that the treatment effect of the warning messages is time-dependent, initially increasing before progressively decreasing over time. Driver responses can be decomposed into several phases at the 95 % confidence level, including reaction time (1.3 s), brake adjustment time (1.3 s), progressive braking duration (2.7 s), and effective treatment duration (4.0 s). The time-dependent bootstrap confidence interval confirms driver heterogeneity in these distinct phases. The proposed functional data analysis approach can serve as a paradigm for quantifying the treatment effect of other ADAS applications. The findings can support the improvements of ADAS design and the development and calibration of driver behavior models accounting for ADAS.</div></div>","PeriodicalId":47520,"journal":{"name":"Analytic Methods in Accident Research","volume":"45 ","pages":"Article 100370"},"PeriodicalIF":12.5,"publicationDate":"2025-01-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143103715","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0