Table 13.

Comparison of A-GRU model accuracy with previous models.

Ref	Model	Acc (%)	Space Complexity	Time Complexity
38	NeuroNet19	99.3	$\mathcal{O}(cwh+1)f$	$\mathcal{O}(f\ast u\ast m)$
39	HHOCNN	98%	$\mathcal{O}(cwh+1)f$	$\mathcal{O}(f\ast u\ast m)$
4	CNN-LSTM	99.22	$\mathcal{O}((nm+mk)+(n\ast d))$	$\mathcal{O}((cwh)+(nd+Kn))$
11	CNN and VGG16	93.36, 97.16	$\mathcal{O}(cwh+1)f$	$\mathcal{O}(f\ast u\ast m)$
14	AE-DNN	96	$\mathcal{O}((n)+(n\ast d))$	$\mathcal{O}((n^2)+(nd+Kn))$
15	ACNN	96.7	$\mathcal{O}(cwh+1)f$	$\mathcal{O}(f\ast u\ast m)$
2024	Proposed model A-GRU	99.32	$\mathcal{O}(cwh+1)f$	$\mathcal{O}(f\ast u\ast m)$

$c=$ the number of convolutional channels, $h=$height of input, $w=$ width of input, $f=$ the convolutional kernel size, $n=$ the number data instances, $k=$ the number of output neurons, $m=$ the number of input neurons and $d=$ the dimension or feature of the input, $K=$number of nearest neighbors, $u=c\ast w\ast h$