Direct multi-step forecasting (ch. 18 p. 548)

One model PER horizon, each trained to predict h steps ahead directly from the same observed lags, so no forecast is ever fed back and errors cannot compound.

Usage

morie_dirmulti(x, lags, horizon)

Arguments

x

the series

lags

positive integer lags

horizon

steps to forecast

Value

list(forecast, horizon, nmodels, first, last, mean)

Examples

morie_dirmulti(x = c(1, 2, 3, 4, 5, 6, 7, 8), lags = 3L, horizon = 5L)
#> $forecast
#> [1]  9 10 11 12 28
#> 
#> $horizon
#> [1] 5
#> 
#> $nmodels
#> [1] 5
#> 
#> $first
#> [1] 9
#> 
#> $last
#> [1] 28
#> 
#> $mean
#> [1] 14
#>