. 2020 Jul 16;15(7):e0233855. doi: 10.1371/journal.pone.0233855

Table 3. Performance of the ARMA forecast models after 2 and 30 weeks in South Korea from the 41^st week of 2014 to the 52^nd week of 2019.

ARMAX	Forecast week	Mean AIC	R²	RMSE	Input variables
ARMAX	Forecast week	Mean AIC	R²	RMSE	Previous seasonal ILI	Total GT	Average vapor pressure	Total INF in Argentina
Reference	2 weeks	917.5	0.623	9.4	O
Reference	30 weeks	743.7	0.247	13.3	O
Forecast	2 weeks	849.1	0.789	7.1	O	O
Forecast	30 weeks	696.0	0.621	9.5	O		O	O

AIC, Akaike Information Criterion; R², Coefficient of determination; RMSE, Root-mean-square error; ARMAX, Auto Regressive Moving Average including exogenous variables; ILI, Influenza-like illness; GT, Google Trends; INF, Influenza

Table 3. Performance of the ARMA forecast models after 2 and 30 weeks in South Korea from the 41st week of 2014 to the 52nd week of 2019.

Table 3. Performance of the ARMA forecast models after 2 and 30 weeks in South Korea from the 41^st week of 2014 to the 52^nd week of 2019.