Bayesian Logistic Regression Models for Credit Scoring by Gregg Webster

Appendix B: R output for logistic regression on “old” data 
Call: 
glm(formula = BAD ~ ., family = binomial, data = old_data) 
Deviance Residuals:
Min 1Q Median 3Q Max
-2.4270 -0.5444 -0.3238 -0.1186 4.0551
Coefficients: 
Estimate Std. Error z value Pr(>|z|)
(Intercept) -7.194e+00 5.636e-01 -12.765 < 2e-16 *** 
LOAN -2.367e-05 6.501e-06 -3.642 0.000271 *** 
MORTDUE -3.710e-06 2.284e-06 -1.625 0.104238
VALUE 3.034e-06 1.596e-06 1.902 0.057212 .
REASONHomeImp 2.028e-01 1.349e-01 1.504 0.132632
JOBOffice -6.819e-01 2.245e-01 -3.038 0.002382 **
JOBOther 1.723e-02 1.786e-01 0.096 0.923139
JOBProfExe 4.760e-02 2.099e-01 0.227 0.820586
JOBSales 4.024e-01 4.245e-01 0.948 0.343111
JOBSelf 4.016e-01 3.799e-01 1.057 0.290496
YOJ -1.615e-02 9.136e-03 -1.768 0.077093 .
DEROG 7.335e-01 8.062e-02 9.098 < 2e-16 *** 

---

117 
DELINQ 8.040e-01 6.416e-02 12.530 < 2e-16 *** 
CLAGE -5.223e-03 8.650e-04 -6.038 1.56e-09 *** 
NINQ 1.367e-01 3.199e-02 4.272 1.94e-05 *** 
CLNO -2.815e-02 6.787e-03 -4.148 3.36e-05 *** 
DEBTINC 1.911e-01 1.378e-02 13.868 < 2e-16 *** 
--- 
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
(Dispersion parameter for binomial family taken to be 1) 
Null deviance: 2797.4 on 2758 degrees of freedom 
Residual deviance: 1866.7 on 2742 degrees of freedom 
AIC: 1900.7 
Number of Fisher Scoring iterations: 6 
Appendix C: R output for logistic regressions on “new” data 
C1: Model with all “new” data 
 
Call: 
glm(formula = BAD ~ ., family = binomial, data = new_data) 
Deviance Residuals:
Min 1Q Median 3Q Max
-2.63880 -0.48115 -0.27481 -0.07361 3.76582
Coefficients: 
Estimate Std. Error z value Pr(>|z|)
(Intercept) -8.615e+00 1.364e+00 -6.317 2.67e-10 *** 
LOAN 5.850e-06 1.382e-05 0.423 0.672030
MORTDUE -6.497e-06 6.965e-06 -0.933 0.350944

---

118 
VALUE 1.618e-06 5.845e-06 0.277 0.781878
REASONHomeImp 1.087e-01 3.223e-01 0.337 0.736028
JOBOffice -9.802e-01 5.819e-01 -1.684 0.092110 .
JOBOther 1.622e-01 4.551e-01 0.357 0.721458
JOBProfExe 1.056e-01 5.287e-01 0.200 0.841722
JOBSales 3.329e+00 9.417e-01 3.535 0.000408 *** 
JOBSelf -1.441e-01 9.040e-01 -0.159 0.873381
YOJ -2.675e-02 2.150e-02 -1.244 0.213402
DEROG 6.563e-01 2.100e-01 3.125 0.001779 **
DELINQ 1.157e+00 1.675e-01 6.904 5.05e-12 *** 
CLAGE -6.654e-03 2.082e-03 -3.196 0.001394 **
NINQ 2.056e-01 6.585e-02 3.122 0.001798 **
CLNO -4.076e-02 1.629e-02 -2.501 0.012374 *
DEBTINC 2.327e-01 3.331e-02 6.985 2.86e-12 *** 
--- 
Signif. codes: 0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1
(Dispersion parameter for binomial family taken to be 1) 
Null deviance: 568.67 on 565 degrees of freedom 
Residual deviance: 341.18 on 549 degrees of freedom 
AIC: 375.18 
Number of Fisher Scoring iterations: 6