Skip to main content

Table 26 Discrimination performance of different feature sets

From: Clues from networks: quantifying relational risk for credit risk evaluation of SMEs

Model

Metric

BF

BF + RS

BF + CN

BF + DN

BF + SN

LR

AUC

0.715 ± 0.015

0.726 ± 0.014

0.722 ± 0.014

0.729 ± 0.014

0.725 ± 0.014

KS

0.408 ± 0.022

0.419 ± 0.022

0.414 ± 0.022

0.431 ± 0.022

0.417 ± 0.021

H

0.251 ± 0.020

0.270 ± 0.020

0.266 ± 0.020

0.283 ± 0.021

0.269 ± 0.020

RF

AUC

0.725 ± 0.014

0.738 ± 0.013

0.738 ± 0.013

0.741 ± 0.013

0.735 ± 0.013

KS

0.426 ± 0.023

0.445 ± 0.020

0.444 ± 0.020

0.450 ± 0.021

0.439 ± 0.022

H

0.265 ± 0.020

0.279 ± 0.018

0.289 ± 0.018

0.294 ± 0.019

0.281 ± 0.019

XGB

AUC

0.695 ± 0.013

0.715 ± 0.013

0.718 ± 0.014

0.722 ± 0.012

0.720 ± 0.014

KS

0.387 ± 0.021

0.416 ± 0.022

0.415 ± 0.021

0.424 ± 0.018

0.434 ± 0.022

H

0.219 ± 0.018

0.252 ± 0.019

0.253 ± 0.019

0.245 ± 0.017

0.257 ± 0.021

  1. Each metric is measured in mean of 100 estimates obtained by repeating the outer ten-fold cross-validation procedure ten times (mean ± standard deviation from a 95% confidence interval)