A method for simplifying the analysis of traffic accidents injury severity on two-lane highways using Bayesian networks

TRYSE Research Group, Department of Civil Engineering, University of Granada, Spain.
Journal of safety research (Impact Factor: 1.34). 10/2011; 42(5):317-26. DOI: 10.1016/j.jsr.2011.06.010
Source: PubMed

ABSTRACT This study describes a method for reducing the number of variables frequently considered in modeling the severity of traffic accidents. The method's efficiency is assessed by constructing Bayesian networks (BN).
It is based on a two stage selection process. Several variable selection algorithms, commonly used in data mining, are applied in order to select subsets of variables. BNs are built using the selected subsets and their performance is compared with the original BN (with all the variables) using five indicators. The BNs that improve the indicators' values are further analyzed for identifying the most significant variables (accident type, age, atmospheric factors, gender, lighting, number of injured, and occupant involved). A new BN is built using these variables, where the results of the indicators indicate, in most of the cases, a statistically significant improvement with respect to the original BN.
It is possible to reduce the number of variables used to model traffic accidents injury severity through BNs without reducing the performance of the model.
The study provides the safety analysts a methodology that could be used to minimize the number of variables used in order to determine efficiently the injury severity of traffic accidents without reducing the performance of the model.

  • [Show abstract] [Hide abstract]
    ABSTRACT: In this paper, we describe a framework for an expert system that tries to predict effects of an accident based on past data using supervised learning employing artificial neural networks. For this purpose, sensory data events are post processed in order to generate a reasonable mapping between input and output parameters in case an event is detected automatically or manually. The framework is intended to be used to take actions for reducing the effects of the accident on traffic congestion and to inform necessary parties to intervene in a timely fashion.
    Computational Intelligence and Informatics (CINTI), 2012 IEEE 13th International Symposium on; 01/2012
  • [Show abstract] [Hide abstract]
    ABSTRACT: Mitigating crash severity on regional transportation roadways is an important concern in road safety research. This paper presents a comprehensive geospatial approach based on the fuzzy classification and regression tree (FCART) to predict motor vehicle crashes and their severity on two-lane, two-way roads. The combined use of fuzzy and decision tree in FCART model solves the uncertainty associated with input data; the model can be easily understood and interpreted because of its graphical tree structure. The FCART model uses fuzzy logic to resolve the difficulty of analyzing input variables where no definitive boundary exists between the categories. Moreover, a bagging algorithm is applied in the FCART model to deal with high-variance crash data and improve the performance of the learning process. The bagged-FCART algorithm is tested against FCART, the classification and regression tree (CART), and the support vector machine (SVM) as inferential engines to predict crash severity and uncover spatial and nonspatial factors that systematically relate to crash severity. The results show that applying the bagging algorithm in the FCART model considerably improves the prediction accuracy and that the bagged-FCART model is superior to other tested models in predicting crash severity. A sensitivity analysis was also conducted to determine the importance of input factors. Parts of the results obtained from this analysis are consistent with the existing traffic safety literature and demonstrate that vehicle failure, drivers wearing seat belts, and weather condition factors are some of the most important factors contributing to crash severity. The proposed approach illustrates that in addition to these factors, geographical factors such as proximity to curves and adjacent facilities and land use have a significant effect on crash severity. These results support the prioritization of effective safety measures that are geographically targeted and behaviorally sound on two-lane, two-way roads.
    Journal of Computing in Civil Engineering 09/2014; 28(5). DOI:10.1061/(ASCE)CP.1943-5487.0000432 · 1.39 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Despite a considerable amount of research on traffic injury severities, relatively little is known about the factors influencing traffic injury severity in developing countries, and in particular in Bangladesh. Road traffic crashes are a common headline in daily newspapers of Bangladesh. It has also recorded one of the highest road fatality rates in the world. This research identifies significant factors contributing to traffic injury severity in Dhaka – a mega city and capital of Bangladesh. Road traffic crash data of 5 years from 2007 to 2011 were collected from the Dhaka Metropolitan Police (DMP), which included about 2714 traffic crashes. The severity level of these crashes was documented in a 4-point ordinal scale: no injury (property damage), minor injury, severe injury, and death. An ordered Probit regression model has been estimated to identify factors contributing to injury severities. Results show that night time influence is associated with a higher level injury severity as is for individuals involved in single vehicle crashes. Crashes on highway sections within the city are found to be more injurious than crashes along the arterial and feeder roads. There is a lower likelihood of injury severities, however, if the road sections are monitored and enforced by the traffic police. The likelihood of injuries is lower on two-way traffic arrangements than one-way, and at four-legged intersections and roundabouts compare to road segments. The findings are compared with those from developed countries and the implications of this research are discussed in terms of policy settings for developing countries.
    4th Road Safety International Conference, Sydney, Australia; 03/2013


Available from
May 31, 2014