Effective Utilization of Data for Predicting COVID-19 Dynamics: An Exploration through Machine Learning Models

Table 2

MAE values of logistic regression models for fatal cases.

Forecast period (days)	Cumulative cases				Daily cases
	Germany	Japan	South Korea	Ukraine	Germany	Japan	South Korea	Ukraine
Train 3	10594.46	4389.47	779.36	2137.49	93.7	37.72	43.5	186.8
Test 3	875	7401	60.33	221.33	68	59.33	7	22.67
Train 7	14431.7	5036.62	1455.71	4339.38	96.37	41.63	41.45	186.37
Test 7	1046.57	5552.86	129.57	511.43	53.29	67.86	14.43	27.29
Train 14	17803.87	6076.29	2126.05	6719.15	102.1	46	41.38	187.03
Test 14	1252.86	5299.07	275.21	1036.36	55.43	78.07	19.43	31.79
Train 21	18747.84	6258.6	2283.06	8983.12	109.16	49.14	43.62	183.2
Test 21	1413.81	6149.38	453.9	1665.52	45.05	76.19	21.86	44.38
Train 30	20143.5	6774.49	2284.55	11098.24	115.11	48.62	45.96	185.85
Test 30	1567.74	7479.45	762.58	2716.65	45.1	83.74	24.1	57.29