Table 4. Comparison among LSTM, Bi-LSTM, GRU, Stacked_Bi_GRU, CNN_LSTM, Deep_CNN, and IndRNN in terms of MAPE.

Prediction of cumulative confirmed COVID-19 cases in India.

Data	Model	Training time(s)	Split (f, t)	MAPE (%)		Fine-tuning time(s)
				No-fine-tuning	Fine-tuning
Confirmed cases	LSTM	733.38	1, 5	11.69	11.03	21.40
			3, 3	17.13	12.07	47.19
			5, 1	25.60	13.24	77.71
	Bi-LSTM	1296.12	1, 5	11.72	11.10	28.82
			3, 3	17.19	11.10	68.27
			5, 1	25.72	11.43	105.57
	GRU	581.84	1, 5	11.72	11.05	20.14
			3, 3	17.18	12.86	45.32
			5, 1	25.70	14.79	70.85
	Stacked_ Bi_GRU	5291.32	1, 5	17.02	15.33	69.52
			3, 3	23.70	19.22	183.36
			5, 1	33.01	24.12	287.70
	CNN_LSTM	3006.03	1, 5	17.58	15.29	41.15
			3, 3	23.88	18.45	104.69
			5, 1	33.16	21.32	170.93
	Deep_CNN	261.37	1, 5	11.63	13.22	13.11
			3, 3	17.21	11.41	29.44
			5, 1	25.84	12.05	45.71
	IndRNN	248.72	1, 5	9.74	9.11	14.85
			3, 3	14.17	8.85	34.96
			5, 1	21.40	0.46	57.08
Confirmed deaths	LSTM	413.85	1, 5	5.29	5.02	20.93
			3, 3	7.94	5.19	48.49
			5, 1	13.21	6.77	74.72
	Bi-LSTM	747.53	1, 5	5.20	5.08	29.68
			3, 3	7.89	5.09	68.36
			5, 1	13.21	6.16	104.81
	GRU	333.79	1, 5	5.21	4.91	19.65
			3, 3	7.80	4.49	44.93
			5, 1	12.96	5.14	69.41
	Stacked_ Bi_GRU	2976.06	1, 5	7.82	7.13	68.48
			3, 3	11.30	9.17	166.58
			5, 1	17.40	11.90	274.43
	CNN_LSTM	1703.61	1, 5	9.49	7.28	41.11
			3, 3	12.91	9.55	100.99
			5, 1	18.90	13.45	173.57
	Deep_CNN	143.17	1, 5	5.62	6.20	13.03
			3, 3	8.60	7.45	29.32
			5, 1	14.37	9.21	44.42
	IndRNN	142.34	1, 5	4.72	4.36	14.54
			3, 3	6.89	3.42	33.95
			5, 1	11.22	2.99	56.54