Choice of refractive surgery types for myopia assisted by machine learning based on doctors’ surgical selection data

Table 3 Accuracy and Macro_F1 of machine learning models on left eye data with different hyperparameters max_depth and with/without feature selection

	Model	Max_depth
		9		10		11
		ACC	Macro_F1	ACC	Macro_F1	ACC	Macro_F1
Feature selection (Select the top 12 features with importance greater than 1.4%.)	NBM	0.7566	0.5696	0.7566	0.5696	0.7566	0.5696
	DBN	0.8276	0.5252	0.8276	0.5252	0.8276	0.5252
	RF	0.8676	0.7635	0.8775	0.8019	0.8725	0.7778
	AdaBoost	0.8374	0.6970	0.8424	0.6871	0.8079	0.6514
	XGBoost	0.8676	0.7460	0.8677	0.7484	0.8725	0.7603
	BP Neural Network	0.8578	0.7462	0.8578	0.7462	0.8578	0.7462
No feature selection was performed (There are 20 features in total.)	NBM	0.6946	0.5643	0.6946	0.5643	0.6946	0.5643
	DBN	0.8054	0.5165	0.8054	0.5165	0.8054	0.5165
	RF	0.8172	0.7148	0.8173	0.6998	0.8226	0.7117
	AdaBoost	0.7784	0.6348	0.7622	0.6284	0.7514	0.6191
	XGBoost	0.8118	0.6872	0.8172	0.6998	0.8226	0.7117
	BP Neural Network	0.7957	0.6271	0.7957	0.6271	0.7957	0.6271

ISSN: 1472-6947