Table 2.

Evaluation results of the prediction models in the Korea National Health and Nutrition Examination Survey dataset.

Feature Selection	Classifier	Accuracy	Sensitivity	Specificity	Precision	F-Score
SVM-RFE	LR	0.7948	0.7818	0.7532	0.7676	0.7746
	RF	0.7890	0.7989	0.7984	0.8115	0.8052
	KNN	0.7342	0.6958	0.7381	0.7961	0.7426
	MLP	0.8070	0.7936	0.7791	0.8016	0.7976
	NN	0.8197	0.8274	0.8203	0.8387	0.8330
	XGBoost	0.8098	0.8108	0.8310	0.8533	0.8315
RFFS	LR	0.7804	0.7371	0.7422	0.8024	0.7684
	RF	0.8264	0.7699	0.7338	0.8236	0.7958
	KNN	0.8048	0.7128	0.7661	0.7753	0.7427
	MLP	0.7994	0.7808	0.7396	0.8115	0.7959
	NN	0.8507	0.8871	0.8902	0.8522	0.8693
	XGBoost	0.8311	0.8782	0.7984	0.8626	0.8703
HFS	LR	0.7834	0.7989	0.7813	0.7959	0.7974
	RF	0.8362	0.7805	0.8496	0.8115	0.7957
	KNN	0.8032	0.8018	0.7123	0.7872	0.7944
	MLP	0.8421	0.8305	0.7513	0.8257	0.8281
	NN	0.8758	0.8518	0.8158	0.8691	0.8604
	XGBoost	0.8812	0.8677	0.8126	0.8737	0.8707

SVM-RFE: support vector machine recursive feature elimination; RFFS: random forest feature selection; HFS: hybrid feature selection; LR: logistic regression; KNN: k-nearest neighbors; NN: neural network; RF: random forest; MLP: multilayer perceptron; XGBoost: extreme gradient boosting. Highest scores are marked in bold.