Table 4 Differences in performance compared with the baseline model.

From: Personalized prediction of delayed graft function for recipients of deceased donor kidney transplants with machine learning

Model

∆Brier Score

IDI

∆PR-AUC

∆ROC-AUC

DeLong test

LR

 − 0.012

 + 0.011

 + 0.006

 + 0.021

p = 0.056

RF

 − 0.013

 + 0.017

 + 0.011

 + 0.026

p = 0.018

EN

 − 0.012

 + 0.011

 + 0.005

 + 0.021

p = 0.056

XGB

 − 0.015

 + 0.025

 + 0.033

 + 0.031

p = 0.005

ANN

 − 0.013

 + 0.018

 + 0.012

 + 0.027

p = 0.012

  1. IDI integrated discrimination improvement, PR-AUC area under the precision-recall curve, ROC-AUC area under the receiver operating characteristic curve, SD standard deviation, BL baseline, LR logistic regression, EN elastic net, RF random forest, XGB extreme gradient boosting, ANN artificial neural network.
Back to article page