Canadian Strategic Highway Research Program
C-SHRP Bayesian Modelling:
A User's Guide


Executive Summary

CHAPTER 1 - INTRODUCTION
The purpose and objectives of the guide are introduced. The impetus and objective for the C-SHRP Bayesian project are discussed.

CHAPTER 2 - 10 STEP TEMPLATE: MODEL DESIGN
The 10-step template for Bayesian regression is introduced. Factors to consider in the design of a Bayesian regression model are discussed. Design considerations include model purpose, dependent and independent variables, and model functional form (Template Steps 1 -5).

CHAPTER 3 - 10 STEP TEMPLATE: DATA AND ANALYSIS
Potential methods for developing a prior are introduced. A step-by-step procedure for developing a prior is presented. An overview of performing Bayesian regression using the available software is presented (Template Steps 6-7).

CHAPTER 4 - 10 STEP TEMPLATE: MODEL EVALUATION
The evaluation and interpretation of the results of Bayesian regression are discussed. This includes comparison of the various priors and evaluation of model performance. The need for further iterations of the model is discussed (Template Steps 8-10).

APPENDIX A - INTRODUCTION TO BAYES' THEOREM
An example of the use of Bayes' theorem is presented as an introductory example. A simple Bayesian regression model is used to introduce key Bayesian regression concepts.

APPENDIX B - BAYESIAN REGRESSION THEORY
The key equations for classical (i.e. non-Bayesian) regression are introduced in matrix notation. The essential theory for Bayesian regression is presented in analogous notation.