Analyzing the severity of accidents on the German Autobahn
Department of Social and Economic Statistics, University of Cologne, Germany. Electronic address: .Accident; analysis and prevention (Impact Factor: 1.65). 04/2013; 57C(8):40-48. DOI: 10.1016/j.aap.2013.03.022
We study the severity of accidents on the German Autobahn in the state of North Rhine-Westphalia using data for the years 2009 until 2011. We use a multinomial logit model to identify statistically relevant factors explaining the severity of the most severe injury, which is classified into the four classes fatal, severe injury, light injury and property damage. Furthermore, to account for unobserved heterogeneity we use a random parameter model. We study the effect of a number of factors including traffic information, road conditions, type of accidents, speed limits, presence of intelligent traffic control systems, age and gender of the driver and location of the accident. Our findings are in line with studies in different settings and indicate that accidents during daylight and at interchanges or construction sites are less severe in general. Accidents caused by the collision with roadside objects, involving pedestrians and motorcycles, or caused by bad sight conditions tend to be more severe. We discuss the measures of the 2011 German traffic safety programm in the light of our results.
[Show abstract] [Hide abstract]
- "The results of inference indicated that the following values of variables were found to be associated with higher probabilities of accidents with casualties including killed or severe injuries: single vehicles accidents, two way undivided roads, fall off vehicle, fixed object collisions, collision with animals, accidents occurring during dark lighting conditions, pavement surface covered with mud, sand or oil, and speed limit higher than 70 km/h. In general, these results were consistent with the literature (Haleem and Gan, 2011; Theofilatos et al., 2012; Manner and Wünsch-Ziegler, 2013; Hosseinpour et al., 2014). Imbalanced data sets can be used to extract important knowledge when balanced, and since sampling techniques have proved their effectiveness in different research areas, this work indicated that they could also be applied in the domain of traffic accidents when used in conjunction with BNs. "
ABSTRACT: Traffic accidents data sets are usually imbalanced, where the number of instances classified under the killed or severe injuries class (minority) is much lower than those classified under the slight injuries class (majority). This, however, supposes a challenging problem for classification algorithms and may cause obtaining a model that well cover the slight injuries instances whereas the killed or severe injuries instances are misclassified frequently. Based on traffic accidents data collected on urban and suburban roads in Jordan for three years (2009-2011); three different data balancing techniques were used: under-sampling which removes some instances of the majority class, oversampling which creates new instances of the minority class and a mix technique that combines both. In addition, different Bayes classifiers were compared for the different imbalanced and balanced data sets: Averaged One-Dependence Estimators, Weightily Average One-Dependence Estimators, and Bayesian networks in order to identify factors that affect the severity of an accident. The results indicated that using the balanced data sets, especially those created using oversampling techniques, with Bayesian networks improved classifying a traffic accident according to its severity and reduced the misclassification of killed and severe injuries instances. On the other hand, the following variables were found to contribute to the occurrence of a killed causality or a severe injury in a traffic accident: number of vehicles involved, accident pattern, number of directions, accident type, lighting, surface condition, and speed limit. This work, to the knowledge of the authors, is the first that aims at analyzing historical data records for traffic accidents occurring in Jordan and the first to apply balancing techniques to analyze injury severity of traffic accidents.
[Show abstract] [Hide abstract]
- "In the area of accident severity research, continuous efforts have been conducted in order to investigate the relationship between the level of severity (dependent variable) and a set of explanatory variables, which usually include: driver attributes (e.g., age and gender), vehicle features (e.g., body type, vehicle age and number of vehicles involved in the accident), road characteristics (e.g., number of lanes, road surface conditions, intersection control and types of road), and accident characteristics (e.g., accident's main cause). Occasionally, the influence of other variables on accident severity like speed limit, day of the week, time of the day, average traffic characteristics (AADT), weather and traffic conditions have also been scrutinized (Delen et al., 2006; Manner and Ziegler, 2013; Torrão, 2013). From previous studies, it is worth mentioning a recent study carried out by Christoforou et al. (2010), mainly because it offers a comprehensive literature review on the subject. "
ABSTRACT: The ordered probit model is used to examine the contribution of several factors to the injury severity faced by motor-vehicle occupants involved in road accidents. The estimated results suggest that motor-vehicle occupants travelling in light-vehicles, at two-way roads, and on dry road surfaces tend to suffer more severe injuries than those who travel in heavy-vehicles, at one-way roads, and on wet road surfaces. Additionally, the driver's seat is clearly the safest seating position, urban areas seem to originate less serious accidents than rural areas, and women tend to be more likely to suffer serious or fatal injuries than men.
[Show abstract] [Hide abstract]
- "Both ordered and unordered models have their own unique benefits and limitations, and the choice of one method over the other is governed by the availability and characteristics of the data and involves taking tradeoffs into consideration (Savolainen et al., 2011; Ye and Lord, 2014). The mixed logit method has been widely applied to analyze crash severities (Anastasopoulos and Mannering, 2011; Chen and Chen, 2011; Haleem and Gan, 2013; Islam et al., 2014; Kim et al., 2008, 2013; Malyshkina and Mannering, 2008; Manner and Wünsch-Ziegler, 2013; Milton et al., 2008; Moore et al., 2011; Morgan and Mannering, 2011; Shaheed et al., 2013; Ye and Lord, 2014; Weiss et al., 2014). The benefit of using this method is that it can accommodate individual unobserved heterogeneity by allowing parameters to differ across observations, and thus, can provide more reliable parameter estimates. "
ABSTRACT: While there have been many studies analyzing crash severity, few studies have accounted for unobserved heterogeneity and compared different crash severity models. The objective of this paper is to investigate the differences between two preferred methods for accommodating individual unobserved heterogeneity, the mixed logit and latent class methods, in exploring the relationship between heavy truck crash severity and its contributing factors. To achieve this, a large sample of crash data on multiple vehicle crashes involving a heavy truck on public roadways in Iowa from 2007 to 2012 was collected. The comparison of the two methods lied on model fit, inferences, and predicted crash severity outcome probabilities. The results suggested a slight superiority of the latent class method in terms of model fit; however, the mixed logit predicted probabilities for all three levels of injury severities were closer (on average) to the observations than the ones predicted by the latent class model. Only a few notable differences in the inferences were found between the two models.
Data provided are for informational purposes only. Although carefully collected, accuracy cannot be guaranteed. The impact factor represents a rough estimation of the journal's impact factor and does not reflect the actual current impact factor. Publisher conditions are provided by RoMEO. Differing provisions from the publisher's actual policy or licence agreement may be applicable.