Bayesian Logistic Regression Models for Credit Scoring by Gregg Webster

2.4 Studies on Bayesian Logistic Regression for Credit Scoring
……………………….....7
 
 
Chapter 3: Methodology and Theoretical Considerations 
3.1 Methodology
…………………………………………………………………………….10
 
3.2 Bayesian Statistics
………………………………………………………………………12
 
3.2.1 Bayesian inference
…………………………………………………………………….12
 
3.2.2 Prior density, likelihood and posterior density functions
…………………………..12
 
3.2.3 Prior distributions
…………………………………………………………………….15 

---

6 
3.3 Generalized Linear Models
…………………………………………………………….16 
3.3.1 Introduction
…………………………………………………………………………...16 
3.3.2 Maximum likelihood estimation
……………………………………………………..18 
3.3.3 Diagnostics
…………………………………………………………………………….21 
3.3.4 Variable selection
……………………………………………………………………..23
 
3.3.5 Logistic regression
……………………………………………………………………24 
3.3.6 Bayesian logistic regression
…………………………………………………………..30 
3.4 Monte Carlo Methods
…………………………………………………………………..31 
3.4.1 Monte Carlo simulation
………………………………………………………………31
 
3.4.2 Markov chains
………………………………………………………………………...38 
3.4.3 Markov chain Monte Carlo
…………………………………………………………..45
 
 
Chapter 4: Results 
4.1 Initial Data Analysis
…………………………………………………………………….51 
4.2 Logistic Regression Model on “old” Data
……………………………………………..57
 
4.3 Determining an Optimal Cut-off Probability
…………………………………………64 
4.4 Logistic Regression Model on “new” Data
……………………………………………66