Identification and optimization of relevant factors for chronic kidney disease in abdominal obesity patients by machine learning methods: insights from NHANES 2005–2018

Table 2 Comparison of discrimination characteristics among ten machine learning models

Model name	Training	Recall	Accuracy	F1-Score	MCC	AUC
AdaBoost	Imbalance	0.376325	0.729278	0.374341	0.201620	1.000
CatBoost	Imbalance	0.298587	0.806844	0.399527	0.326193	0.875
Decision Tree	Imbalance	0.287986	0.806844	0.390887	0.322116	0.754
Gradient Boosting Decision Tree	Imbalance	0.362191	0.730798	0.366726	0.195853	1.000
LightGBM	Imbalance	0.280919	0.800000	0.376777	0.298428	0.944
Logistic regression	Imbalance	0.266784	0.806844	0.37284	0.314087	0.768
Naive Bayes	Imbalance	0.613074	0.660456	0.437303	0.241620	0.714
Random forest	Imbalance	0.286219	0.805703	0.388024	0.317875	0.780
Support vetor machine	Imbalance	0.000000	0.784791	NA	NA	0.646
XGBoost	Imbalance	0.323322	0.801521	0.412162	0.320935	0.990
AdaBoost	Balance	0.887827	0.848218	0.875550	0.681574	0.838
CatBoost	Balance	0.886552	0.867382	0.889386	0.723863	0.938
Decision Tree	Balance	0.768005	0.793024	0.816949	0.587348	0.863
Gradient Boosting Decision Tree	Balance	0.891013	0.853200	0.879522	0.692142	0.869
LightGBM	Balance	0.852135	0.728248	0.790423	0.418553	0.783
Logistic regression	Balance	0.873168	0.811422	0.847772	0.602344	0.886
Naive Bayes	Balance	0.698534	0.666922	0.716106	0.314429	0.823
Random Forest	Balance	0.787126	0.803756	0.828303	0.605274	0.891
Support vetor machine	Balance	0.876992	0.804140	0.843396	0.586093	0.882
XGBoost	Balance	0.825420	0.846691	0.843834	0.694092	0.923

Abbreviations: MCC Matthews correlation coefficient, AUC Area under the receiver operator curve

ISSN: 1476-511X