Table 2.

The classification performance of different convolutional algorithms

Cases	Parameters	Various convolutional algorithms
		Conventional CNN	Attention-based CNN	Proposed hierarchical Attention-based CNN
Case I	Accuracy (%)	96.19	97.2826	98.33
	Sensitivity	0.9474	0.9780	0.9800
	Specificity	0.9775	0.9677	0.9700
	Precision	0.9783	0.9674	0.9780
	F-measure	0.9626	0.9727	0.9800
	MCC	0.9244	0.9457	0.9600
Case II	Accuracy (%)	95.11	95.11	95.56
	Sensitivity	0.9368	0.9462	0.9667
	Specificity	0.9663	0.9560	0.9444
	Precision	0.9674	0.9565	0.9457
	F-measure	0.9519	0.9514	0.9560
	MCC	0.9027	0.9022	0.9113
Case III	Accuracy (%)	95.65	96.20	97.21
	Sensitivity	0.9375	0.9570	0.9775
	Specificity	0.9773	0.9670	0.9667
	Precision	0.9783	0.9674	0.9667
	F-measure	0.9574	0.9622	0.9721
	MCC	0.9139	0.9240	0.9442