Bayesian Logistic Regression Models for Credit Scoring by Gregg Webster

 
4.5 Bayesian Logistic Regression Model on “new” Data
…………………………………73
 
4.6 Performance of Models on Test Data
………………………………………………….81 
4.6.1 Cut-off probability of 0.3
……………………………………………………………..82 
4.6.2 Cut-off probability of 0.48
……………………………………………………………84 
4.6.3 Comparison of the two cut-off probabilities
………………………………………...87
 
4.7 Performance of the Models with Varying Amounts of “new” Data
…………………87 

---

7 
4.7.1 Cut-off probability of 0.3
……………………………………………………………..88 
4.7.2 Cut-off probability of 0.48
……………………………………………………………89
 
4.8 Conclusions
……………………………………………………………………………...91
 
 
Chapter 5: Conclusions and Implications 
5.1 Summary
………………………………………………………………………………...93
 
5.2 Limitations, Recommendations and Further Research
………………………………94
 
 
References
…………………………………………………………………………………...96 
Appendix
…………………………………………………………………………………...100 
 
 
 

---

8 
LIST OF FIGURES 
 
Fig. 4.1 
Bar plots of REASON and JOB……………………………………………………..53 
Fig. 4.2 
Histogram and Box plot of LOAN………………………………………………….53 
Fig. 4.3
Histogram and Box plot of MORTDUE…………………………………………….54 
Fig. 4.4
Histogram and Box plot of VALUE………………………………………………...54 
Fig. 4.5
Histogram and Box plot of DEBTINC……………………………………………...54 
Fig. 4.6
Histogram and Box plot of YOJ…………………………………………………….55 
Fig. 4.7
Histogram and Box plot of DEROG………………………………………………...55