Table 3.

The average classification performance of different algorithms

Parameters	Various convolutional algorithms
	Conventional CNN	Attention-based CNN	Proposed hierarchical Attention-based CNN
Accuracy (%)	95.65	96.20	97.03
Sensitivity	0.9406	0.9604	0.9747
Specificity	0.9737	0.9636	0.9604
Precision	0.9746	0.9638	0.9634
F-measure	0.9573	0.9621	0.9694
MCC	0.9137	0.9240	0.9385