Skip to main content

Table 1 MAPEs for in-sample and out-of-sample approach in the two different forecast period

From: Exploring different modelling approaches to forecast the community acute respiratory infections burden in children: an Italian epidemiological time series study

Model

In-sample

Out-of-sample

Overall

A) September 2010—February 2020

 ETS(M,N,M)

6.93

5.43

6.65

 TBATS(0.485,{0,0},-,{< 12,5 >}

7.11

5.61

7.77

 ARIMA(1,0,0)(1,1,0) [12] with drifta

14.60

19.98

17.62

 UCM with irregular, level, slope and seasonality modelled with a fourth degree spline and 6 knots from December to May

12.27

10.40

10.97

B) September 2010 – September 2023 with forecasted COVID-19

 ETS(M,N,M)

6.69

7.46

6.85

 TBATS(0.123,{0,0},-,{< 12,5 >}

6.72

6.81

6.87

 ARIMA(1,0,0)(2,1,0) [12] with drifta

13.07

29.94

25.33

 UCM with irregular, level, slope and seasonality modelled with a fourth degree spline and 6 knots from December to May

11.70

10.20

11.08

  1. aApplied on log-transformed series