Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

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

Yuanchang Xie1, Dominique Lord, Yunlong Zhang

  • 1Zachry Department of Civil Engineering, Texas A&M University, 3136 TAMU, College Station, TX 77843-3136, United States. ycwie@tamu.edu

Accident; Analysis and Prevention
|February 20, 2007
PubMed
Summary

Bayesian neural network (BNN) models show superior performance in predicting motor vehicle crashes compared to back-propagation neural networks (BPNN) and negative binomial regression. BNNs effectively mitigate data overfitting, enhancing generalization abilities in highway safety analysis.

Related Concept Videos

Mechanistic Models: Compartment Models in Individual and Population Analysis01:23

Mechanistic Models: Compartment Models in Individual and Population Analysis

Mechanistic models are utilized in individual analysis using single-source data, but imperfections arise due to data collection errors, preventing perfect prediction of observed data. The mathematical equation involves known values (Xi), observed concentrations (Ci), measurement errors (εi), model parameters (ϕj), and the related function (ƒi) for i number of values. Different least-squares metrics quantify differences between predicted and observed values. The ordinary least squares (OLS)...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Microstructural Evolution and Mechanical Response of UHPS Incorporating Steel Slag Fine Aggregate and Different Morphology Steel Fibers.

Materials (Basel, Switzerland)·2026
Same author

Transthyretin Tetramer as a Prognostic Biomarker in Heart Failure.

JACC. Asia·2026
Same author

Effects of dietary supplementation with bile acids on production performance, egg quality, nutrient digestibility, and morphology of reproductive tract in late-laying hens.

Poultry science·2026
Same author

Divergent futures for critical ecological areas: How climate pathways reshape the roles of ecosystem integrity, multifunctionality, and stability.

Journal of environmental management·2026
Same author

Breaking the O<sub>2</sub> Activation Bottleneck via Synergistic Electronic Polarization at Adjacent Co-Mn Sites in N-Doped Carbon.

Chemistry (Weinheim an der Bergstrasse, Germany)·2026
Same author

Highly Selective Purification of Trace <sup>131</sup>I in Radioactive Wastewater via Adaptive Inflatable Organic Cages with Fully Accessible Sites.

Inorganic chemistry·2026

Area of Science:

  • Transportation Safety
  • Statistical Modeling
  • Machine Learning in Engineering

Background:

  • Statistical models are crucial for highway safety studies, including variable relationships, covariate screening, and value prediction.
  • Generalized linear models (GLM) and hierarchical Bayes models (HBM) are traditional methods, while back-propagation neural networks (BPNN) are newer but face estimation complexity and overfitting issues.
  • Bayesian neural network (BNN) models have emerged as a solution to overcome BPNN limitations, offering improved performance and reduced overfitting.

Purpose of the Study:

  • To evaluate the application and predictive performance of Bayesian neural network (BNN) models for motor vehicle crash prediction.
  • To compare BNN models against back-propagation neural network (BPNN) and negative binomial (NB) regression models in highway safety contexts.

Related Experiment Videos

Main Methods:

  • Estimation of three distinct models: Bayesian neural network (BNN), back-propagation neural network (BPNN), and negative binomial (NB) regression.
  • Utilized data collected from rural frontage roads in Texas for model estimation and comparison.
  • Performance evaluation focused on data prediction accuracy and data fitting capabilities.

Main Results:

  • Both neural network models (BPNN and BNN) generally outperformed the negative binomial (NB) regression model in data prediction.
  • The Bayesian neural network (BNN) model consistently demonstrated superior data fitting and prediction performance compared to the back-propagation neural network (BPNN) model.
  • BNN models effectively alleviated the overfitting problem inherent in BPNN models without compromising nonlinear approximation abilities, indicating better generalization.

Conclusions:

  • Bayesian neural network (BNN) models are highly effective for predicting motor vehicle crashes and offer significant advantages over traditional and other neural network approaches.
  • BNNs provide enhanced generalization capabilities, making them a valuable tool for mitigating overfitting in highway safety modeling.
  • The study suggests BNNs can be applied to other critical areas in highway safety, such as developing accident modification factors and improving evaluations of highway design alternatives.