An Intelligent Diagnostic System to Analyze Early-Stage Chronic Kidney Disease for Clinical Application

Table 8

Experimental result analysis of different classifiers for the hybrid dataset (mixture of CKD-15 and CKD-21).

Classifier	Recall	TNR	PPV	F1-score	Train accuracy	Test accuracy	ROC-AUC
MLP	0.98	1	1	0.99	100	99.12	0.991
DT	0.96	1	1	0.98	100	97.81	0.978
GB	0.96	1	1	0.98	100	97.81	0.978
RF	0.96	1	1	0.98	99.62	97.81	0.978
SGD	0.97	0.98	0.98	0.97	98.49	97.81	0.978
ADAB	0.95	1	1	0.97	100	97.37	0.974
KNN	0.95	1	1	0.97	98.31	97.37	0.974
SVM	0.96	0.99	0.99	0.97	98.12	97.37	0.974
LR	0.95	0.99	0.99	0.97	98.12	96.93	0.996
ELM	0.97	0.96	0.96	0.96	0.98	96.49	0.985
PAC	0.97	0.96	0.96	0.96	98.31	96.49	0.965
NB	0.9	1	1	0.95	96.99	95.18	0.992

The best results are indicated in bold.