A hybrid stacked ensemble and Kernel SHAP-based model for intelligent cardiotocography classification and interpretability

Table 7 Comparison with several current CTG-based classifier in the public dataset

Classifiers	Categories	F1	Average F1	Accuracy
MLPNN [8]	Normal	0.9500	0.8131	0.9036
	Suspicious	0.6843
	Pathologic	0.8050
GRNN [8]	Normal	0.9570	0.8483	0.9186
	Suspicious	0.7392
	Pathologic	0.8488
PNN [8]	Normal	0.9591	0.8506	0.9214
	Suspicious	0.7381
	Pathologic	0.8545
Random Forest [9]	Normal	0.9730	0.8996	0.9480
	Suspicious	0.8260
	Pathologic	0.9000
DF [7]	Normal	0.9700	0.9201	0.9507
	Suspicious	0.8571
	Pathologic	0.9333
Proposed strategy	Normal	0.9779	0.9345	0.9639
	Suspicious	0.8977
	Pathologic	0.9434

ISSN: 1472-6947