Predictive accuracy of a hierarchical logistic model of cumulative SARS-CoV-2 case growth until May 2020

Table 1 Summary estimates of predictive accuracy

	RMSE (95% CI)	MAPE (95% CI)	R² (95% CI)	ICC (95% CI)
Day of estimation	0.640 (0.577 to 0.707)	0.323 (0.295 to 0.356)	0.989 (0.986 to 0.992)	0.984 (0.979 to 988)
Two weeks forecast	0.900 (0.803 to 1.05)	1.085 (0.673 to 2.598)	0.980 (0.971 to 0.984)	0.935 (0.905 to 0.950)
One month forecast	1.393 (1.271 to 1.546)	2.133 (1.600 to 2.953)	0.958 (0.948 to 0.966)	0.828 (0.777 to 0.866)
Two months forecast	1.958 (1.791 to 2.157)	4.250 (2.907 to 6.735)	0.931 (0.914 to 0.943)	0.679 (0.606 to 0.748)

RMSE root mean squared error in logarithmic case counts, MAPE mean absolute percentage error in case counts, R² coefficient of determination, ICC intraclass correlation, CI confidence interval

ISSN: 1471-2288