Table 2.

Performance of various prediction models.

Model			Training AUCa		Testing AUC		Precision	Recall		F1-score
5-year follow-up
	LRb	0.72		0.67		0.01		0.70	0.02
	LDAc	0.80		0.81		0.02		0.88	0.04
	LGBMd	0.94		0.82		0.02		0.90	0.04
	GBMe	0.86		0.82		0.02		0.95	0.05
	RFf	0.95		0.81		0.02		0.88	0.04
	XGBoostg	0.98		0.80		0.02		0.86	0.04
	AdaBoost	0.82		0.81		0.02		0.85	0.04
	Voting	0.86		0.82		0.02		0.91	0.05
	ANNh,i	0.98		0.97		0.01		0.70	0.03
10-year follow-up
	LR	0.68		0.67		0.01		0.73	0.03
	LDA	0.78		0.78		0.02		0.85	0.04
	LGBM	0.92		0.80		0.02		0.88	0.05
	GBM	0.84		0.80		0.02		0.90	0.05
	RF	0.94		0.80		0.02		0.85	0.05
	XGBoost	0.97		0.78		0.02		0.83	0.05
	AdaBoost	0.81		0.80		0.02		0.84	0.05
	Voting	0.84		0.80		0.02		0.89	0.05
	ANNi	0.98		0.98		0.02		0.79	0.03

^aAUC: area under the curve.

^bLR: logistic regression.

^cLDA: linear discriminant analysis.

^dLGBM: light gradient boosting machine.

^eGBM: gradient boosting machine.

^fRF: random forest.

^gXGBoost: extreme gradient boosting.

^hANN: artificial neural network.

ⁱBest model based on area under the curve values.