Table 3.

Parameter sensitivity analysis of the proposed hybrid deep learning model in terms of the number of transformer heads with three ensemble deep learning options.

No. of Pre-trained Ensemble Models	Transformer Head	ACC	PRE	F1-score
One Model	2	0.9684	0.9612	0.9602
Two Models		0.9695	0.9711	0.9608
Three Models		0.9771	0.9721	0.9729
One Model	4	0.9696	0.9623	0.9609
Two Models		0.9770	0.9801	0.9711
Three Models		0.9820	0.9756	0.9744
One Model	8	0.9699	0.9701	0.9703
Two Models		0.9799	0.9800	0.9708
Three Models		0.9898	0.9806	0.9801