Using fixed-parameter and random-parameter ordered regression models to identify significant factors that affect the severity of drivers' injuries in vehicle-train collisions

Exploring Influential Factors Affecting the Severity of Urban Expressway Collisions: A Study Based on Collision Data

When traffic collisions occur on urban expressways, the consequences, including injuries, the loss of lives, and damage to properties, are more serious. However, the existing research on the severity of expressway traffic collisions has not been deeply explored. The purpose of this research was to investigate how various factors affect the severity of urban expressway collisions. The severity of urban expressway collisions was set as the dependent variable, which could be divided into three categories: slight collisions, severe collisions, and fatal collisions. Ten variables, including individual characteristics, collision characteristics, and road environment conditions, were selected as independent factors. Based on 975 valid urban expressway collisions, an ordered logistic regression model was established to evaluate the impacts of influence factors on the severity of these crashes. The results show that gender, collision modality, road pavement conditions, road surface conditions, and visibility are significant factors that affect the severity of urban expressway collisions. Females were more likely to be involved in more severe urban expressway collisions than males. For collisions involving pedestrians and non-motorized vehicles, the risk of more severe injury was 7.508 times higher than that associated with vehicle-vehicle collisions. The probability of more severe collisions on urban expressways with poor pavement conditions and wet surface conditions is greater than that on urban expressways with good pavement conditions and dry surface conditions. In addition, as visibility increases, the probability of more severe collisions on urban expressways gradually decreases. These results provide more effective strategies to reduce casualties as a result of urban expressway collisions.

Railroad accident analysis using extreme gradient boosting

Railroads are critical to the economic health of a nation. Unfortunately, railroads lose hundreds of millions of dollars from accidents each year. Trends reveal that derailments consistently account for more than 70 % of the U.S. railroad industry's average annual accident cost. Hence, knowledge of explanatory factors that distinguish derailments from other accident types can inform more cost-effective and impactful railroad risk management strategies. Five feature scoring methods, including ANOVA and Gini, agreed that the top four explanatory factors in accident type prediction were track class, type of movement authority, excess speed, and territory signalization. Among 11 different types of machine learning algorithms, the extreme gradient boosting method was most effective at predicting the accident type with an area under the receiver operating curve (AUC) metric of 89 %. Principle component analysis revealed that relative to other accident types, derailments were more strongly associated with lower track classes, non-signalized territories, and movement authorizations within restricted limits. On average, derailments occurred at 16 kph below the speed limit for the track class whereas other accident types occurred at 32 kph below the speed limit. Railroads can use the integrated data preparation, machine learning, and feature ranking framework presented to gain additional insights for managing risk, based on their unique operating environments.

Single-vehicle crash severity outcome prediction and determinant extraction using tree-based and other non-parametric models

Single-vehicle crashes are more fatality-concentrated and have posed increasing challenges in traffic safety, which is of great research necessity. Tremendous previous studies have conducted relevant analysis with econometric modeling approaches, whereas the ability of non-parametric methods to predict crash severity is still smattering of knowledge. Consequently, the main objective of this paper is to conduct single-vehicle crash severity prediction with different tree-based and non-parameter models. An alternate aim is to identify the intrinsic mechanism of how contributing factors determine single-vehicle crash severity. By virtue of Grid-Search method, this paper conducted fine-tuning of different models to obtain the best performances based on five crash severity sub-datasets. For model evaluation, the accuracy indicators were calculated in training, validation and test sets, respectively. Besides, feature importance extraction was undertaken based on the results of model comparison. The finding indicated that these models didn't exhibit a huge performance difference for crash severity prediction in the same severity level; however, the performances of the models did vary among different datasets, with an average training accuracy of 99.27 %, 96.4 %, 86.98 %, 86.84 %, 71.76 % in fatal injury, severe injury, visible injury, complaint of pain, PDO crash datasets, respectively. Additionally, it was found that in each severity dataset, the indicator urban freeways is a determinant factor that leads to the occurrence of crashes while rural freeways is more related to more severe crashes (i.e., fatal and severe crashes). This paper can provide valuable information for model selection and tuning in accident severity prediction. Future research could consider the influences that temporal instability of contributing features has on the model performances.

Investigating hierarchical effects of adaptive signal control system on crash severity using random-parameter ordered regression models incorporating observed heterogeneity

By handling conflicting traffic movements and establishing dynamic coordination between intersections in real-time, the Adaptive Signal Control System (ASCS) can potentially improve the operation and safety of signalized intersections on a corridor. This study identifies the hierarchical effects of ASCS on the crash severity by exploring the heterogeneous effect of ASCS on the crash severity. Four different random-parameter ordered regression models (two ordered probit models, and two ordered logit models) are developed and compared. The analysis reveals that the random-parameter ordered probit and logit models (ROP and ROL) with observed heterogeneity perform better than the random-parameter ordered probit and logit models (RP and RL) without observed heterogeneity in terms of the Akaike information criteria and the goodness of fit of the model. The ROP model performs better than the ROL model in terms of classification model performance measures. The ROP model enables parameters (i.e., the coefficients of the explanatory variables) to vary as a function of explanatory variables as well as across observations, thus accounting for both observed (captured by available explanatory variables) and unobserved (not captured by available explanatory variables) heterogeneity. The analysis reveals that the presence of ASCS is associated with lower crash severity. In this study, observed heterogeneity of ASCS effects on the crash severity is captured by variables related to the intersection and corridor features. Other contributing factors besides ASCS, such as annual average daily traffic, speed limit, lighting, peak period, crash type (rear-end, angle), and pedestrian involvements, are also associated with the probability of crash severity. Unobserved heterogeneity of the effect of angle crash type on the crash severity is found to exist across the observations. The findings of this research have practical implications for establishing ASCS implementation guidelines in lowering the probability of higher crash severity.

Influence of built environment on the severity of vehicle crashes caused by distracted driving: A multi-state comparison

With recent increased attention to the consequences of distracted driving (DD), this research provides a comprehensive investigation of the role of the built environment on the severity of vehicle crashes caused by DD. Utilizing crash data collected from fifteen states in the United States for the period 2013-2017, the association between distracted driving crash severity and various built environment indicators was examined by the generalized ordered logit regression model. The results show that at a lower severity level, DD related crashes were found to be less severe at roundabouts or in urban areas, whereas the probability of injuries rather than property damage only (PDO) increases if an accident involves speeding or when occurring at an intersection or a curved road. Comparatively, at a higher severity level, the odds of severe (or fatal) injury involvement compared to minor injuries and PDO was found to be higher in a work-zone, a curved roadway, or when excessive speed was involved. Conversely, roundabouts and urban areas affected negatively in severe DD crash, which is consistent with the lower-level case. The study also reveals a state-specific variability of the influence of the built environment on the severity of DD related crashes. These findings provide a comprehensive understanding of the severity of DD related crashes for transportation safety planners or policymakers to develop customized policy recommendations, such as designing policies or roadway safety treatments, to curb the negative consequences of distracted driving.

Temporal stability of the factors related to the severity of drivers' injuries in rear-end collisions

Most of the previous studies that investigated the factors increasing the severity of rear-end collisions were based on analyzing collision reports from multiple years and combining them into a single dataset for analysis. Analyzing pooled data from multiple years carries the risk of introducing aggregation bias in the analysis. Those aggregated models might be structurally unstable, and the significance of the risk factors identified using those aggregated models might change over time due to the ongoing changes in vehicle technologies, law-enforcement technologies, and drivers' attitudes. This study demonstrates the importance of testing the temporal stability of pooled data by utilizing logistic regression modeling to analyze all rear-end collisions that occurred in North Carolina for the period from January 1, 2004 to December 31, 2015. Separate models were developed for each year to model injury severity of striking and struck drivers. The year-wise models were compared together to identify the most temporally stable factors, and it was found that older and female drivers are usually more severely injured, but they do not increase injury severity of the drivers they collide with. It was also found that compared to other light-duty vehicles, passenger cars are usually associated with increased injury severity to their drivers and reduced injury severity to the drivers of the vehicles they collide with. The increased age of a vehicle was found to increase the injury severity of its driver as well as the driver of the vehicle it collides with. Dark conditions were found to increase drivers' injury severity, but adverse weather conditions have no similar effect. For comparison, aggregated models were also developed by pooling data from all analysis years (from 2004 to 2015) and were found to return significant factors that were found by the year-wise models to be temporally unstable. Chow tests were performed on the data, and it was found that pooling data for four years or more generally returned structurally unstable models.