Table 6.

Results on three datasets using proposed strategies

Dataset	Pretrained Model	Ensemble	Accuracy	Precision	Recall	F1-score	QWK
Eyepacs	None	No	0.921	0.923	0.923	0.924	0.878
Eyepacs	None	Yes	0.944	0.943	0.942	0.932	0.901
Aptos	None	No	0.942	0.943	0.943	0.944	0.954
Aptos	Eyepacs	No	0.954	0.952	0.942	0.954	0.956
Aptos	Eyepacs	Yes	0.954	0.953	0.95	0.956	0.967
Messidor2	None	No	0.854	0.856	0.853	0.844	0.854
Messidor2	Eyepacs	No	0.923	0.922	0.923	0.912	0.933
Messidor2	Eyepacs	Yes	0.924	0.922	0.923	0.921	0.944