Table 6. The testing results of CNN_3, CNN_4, and HSplice performed on the donor sites of the combined dataset.

The testing results performed on 20% of each hold-out set of the donor sites.

Site	Evaluation Metrics	Model	Ratio
			1:1	1:3	1:5	1:7
Donor	AUPRC	CNN_3	0.795	0.846	0.891	0.935
		CNN_4	0.798	0.847	0.892	0.934
		HSplice	0.551	0.527	0.686	0.594
	Precision	CNN_3	0.394	0.615	0.760	0.889
		CNN_4	0.387	0.561	0.708	0.850
		HSplice	0.290	0.327	0.443	0.539
	Recall	CNN_3	0.960	0.919	0.888	0.854
		CNN_4	0.962	0.940	0.922	0.903
		HSplice	0.927	0.813	0.909	0.660
	MCC	CNN_3	0.599	0.740	0.810	0.856
		CNN_4	0.594	0.713	0.795	0.861
		HSplice	0.496	0.487	0.606	0.542

Notes.

Bold styling emphasizes the highest values regarding the evaluation metrics used in the study.