Predicting motor vehicle collisions using Bayesian neural network models: An empirical analysis

Xie, Yuanchang; Lord, Dominique; Zhang, Yun-Long

doi:10.1016/j.aap.2006.12.014

Cited by 207 publications

(124 citation statements)

References 51 publications

Supporting

Mentioning

117

Contrasting

Unclassified

Order By: Relevance

“…In fact, it is not essential for a researcher to employ a regression model at all in any of the two stages. For instance, one can use a neural network model in the accident frequency analysis (Xie et al, 2007;Lord and Mannering, 2010) and a data mining technique such as the classification and regression tree approach in the severity analysis (Chang and Wang, 2006). This may also benefit the practitioners in that two teams are able to work on the frequency and severity analyses separately and the results can then be combined.…”

Section: Discussionmentioning

confidence: 99%

“…The accuracy of the two-stage model through goodness-of-fit can be determined by a number of statistics such as mean absolute deviation (MAD), and mean squared error (MSE). For example, Oh et al (2003) and Xie et al (2007) employed the MAD statistics and indicated that a lower MAD characterises a better model in term of predicting accuracy. After obtaining the expected number of accidents at each severity level, road segments can then be ranked by an appropriate decision parameter (Θ) for further engineering examination and treatment.…”

Section: Predicting Accident Frequency At Different Severity Levels Amentioning

confidence: 99%

See 1 more Smart Citation

Predicting accident frequency at their severity levels and its application in site ranking using a two-stage mixed multivariate model

Wang

Quddus

Ison

2011

Accident Analysis & Prevention

119

View full text Add to dashboard Cite

Citation: WANG, C., QUDDUS, M.A. and ISON, S.G., Predicting accident frequency at their severity levels and its application in site ranking using a two-stage mixed multivariate model. Accident Analysis and Prevention, 43(6), pp.1979-1990 Metadata AbstractAccident prediction models (APMs) have been extensively used in site ranking with the objective of identifying accident hotspots. Previously this has been achieved by using a univariate count data or a multivariate count data model (e.g. multivariate Poissonlognormal) for modelling the number of accidents at different severity levels simultaneously. This paper proposes an alternative method to estimate accident frequency at different severity levels, namely the two-stage mixed multivariate model which combines both accident frequency and severity models. The accident, traffic and road characteristics data from the M25 motorway and surrounding major roads in England have been collected to demonstrate the use of the two-stage model. A Bayesian spatial model and a mixed logit model have been employed at each stage for accident frequency and severity analysis respectively, and the results combined to produce estimation of the number of accidents at different severity levels. Based on the results from the two-stage model, the accident hotspots on the M25 and surround have been identified. The ranking result using the twostage model has also been compared with other ranking methods, such as the naïve ranking method, multivariate Poisson-lognormal and fixed proportion method. Compared to the traditional frequency based analysis, the two-stage model has the advantage in that it utilises more detailed individual accident level data and is able to predict low frequency accidents (such as fatal accidents). Therefore, the two-stage mixed multivariate model is a promising tool in predicting accident frequency according to their severity levels and site ranking.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Predicting Accident Frequency At Different Severity Levels Amentioning

confidence: 99%

Predicting accident frequency at their severity levels and its application in site ranking using a two-stage mixed multivariate model

Wang

Quddus

Ison

2011

Accident Analysis & Prevention

119

View full text Add to dashboard Cite

show abstract

“…Other researchers, however, consider these models on a class of their own (Gelman and Hill, 2007). models have been used in highway safety (Abdelwahab and Abdel-Aty, 2002;Chang, 2005;Riviere et al, 2006;Xie et al, 2007) mainly as a predictive tool. Overall, these models tend to exhibit better linear/non-linear approximation properties than traditional count-model approaches.…”

mentioning

confidence: 99%

The statistical analysis of crash-frequency data: A review and assessment of methodological alternatives

Lord

Mannering

2010

Transportation Research Part A: Policy and Practice

Self Cite

980

739

View full text Add to dashboard Cite

“…Neural Networks (ANN) (Xie et al 2007;Chang 2005;Abdelwahab and Abdel-Aty 2002), Generalized Additive Models (Li et al 2011;Xie and Zhang 2008), Quintile regression (QR) (Wu et al 2014;Qin and Reyes 2011;Qin et al 2010), Bayesian Methods (Ma et al 2008;Miaou and Song 2005;Pawlovich et al 2006;Washington and Oh 2006), etc. in traffic safety studies.…”

Section: Traffic Crash Predictionmentioning

confidence: 99%

Heterogeneous Data Uncertainties in Risk Management of a Multiscale Transportation Program

View full text Add to dashboard Cite

Access management, which systematically limits opportunities for egress and ingress of vehicles to highway lanes, is critical to protect trillions of dollars of current investment in transportation. While access management can be effective to avoid crashes, reduce travel times, and increase route capacities, the literature suggests a need for metrics to guide investments in resource allocation across large networks at several time horizons and geographic scales. This dissertation describes a decision aid to support a multiscale transportation access management program via a risk-cost-benefit tradeoff analysis with heterogeneous sources of data and expertise, addressing incomplete or partially relevant information on regions, decision criteria, crash rates, travel speeds, road condition, project costs, and other factors. The approach quantifies safety improvement, travel-time savings, and costs of access management through functional relationships of input parameters including crash rates, corridor access point densities, and traffic volumes. Parameter uncertainties, which vary across regions and time horizons, are addressed via numerical interval analyses. The integration of methods is demonstrated for 6,000 highway miles of a 43,000 square-mile region and its several sub-regions. The demonstration prioritizes route segments that would benefit from risk assessment and risk management, including (i) right of way purchases, (ii) restriction of access points, (iii) new alignments, (iv) developer proffers, (v) further data collection, (vi) further expert elicitation, (vii) etc. The philosophy of approach is generally applicable to address uncertainties of heterogeneous data in resource allocation and decision making for multiscale systems.

show abstract

Predicting motor vehicle collisions using Bayesian neural network models: An empirical analysis

Cited by 207 publications

References 51 publications

Predicting accident frequency at their severity levels and its application in site ranking using a two-stage mixed multivariate model

Predicting accident frequency at their severity levels and its application in site ranking using a two-stage mixed multivariate model

The statistical analysis of crash-frequency data: A review and assessment of methodological alternatives

Heterogeneous Data Uncertainties in Risk Management of a Multiscale Transportation Program

Contact Info

Product

Resources

About