Table 5. Performance comparison of various models with different sequence lengths under individual learning strategies.

A performance comparison of various models using different sequence lengths (300 and 600 bp) under individual learning strategies. The metrics include loss, accuracy (Acc), AUC, F1-score, specificity (Sp), and Matthews correlation coefficient (MCC). The results show how models like DNABERT-2_BASE, DNABERT-2_CNN, and DNABERT-2_BiLSTM perform differently across these sequence lengths. The best results for sequence lengths of 300 bp and 600 bp are indicated in bold.

Length	Model	Loss	Acc	AUC	F1	Sp	MCC
300 bp	DNABERT-2_BASE	0.518	0.850	0.895	0.849	0.933	0.710
	DNABERT-2_CNN	0.509	0.858	0.931	0.858	0.925	0.723
	DNABERT-2_BiLSTM	0.467	0.867	0.841	0.865	0.967	0.749
	DNABERT-2_C M _BL	0.395	0.858	0.924	0.857	0.942	0.727
	DNABERT-2_C A _BL	0.495	0.867	0.903	0.866	0.942	0.742
600 bp	DNABERT-2_BASE	0.537	0.846	0.916	0.843	0.975	0.716
	DNABERT-2_CNN	0.567	0.871	0.940	0.870	0.958	0.753
	DNABERT-2_BiLSTM	0.516	0.867	0.909	0.867	0.883	0.734
	DNABERT-2_C M _BL	0.429	0.842	0.931	0.839	0.958	0.703
	DNABERT-2_C A _BL	0.333	0.904	0.949	0.904	0.933	0.810