Applies a fitted model to a new dataset. For most methods this can be done with or without re-estimation of the parameters.
library(fable)
fit <- as_tsibble(mdeaths) %>%
model(ETS(value ~ error("M") + trend("A") + season("A")))
fit %>% report()
#> Series: value
#> Model: ETS(M,A,A)
#> Smoothing parameters:
#> alpha = 0.0002065548
#> beta = 0.0001865257
#> gamma = 0.000118306
#>
#> Initial states:
#> l[0] b[0] s[0] s[-1] s[-2] s[-3] s[-4] s[-5]
#> 1671.676 -4.334248 373.1746 -121.3157 -246.1697 -484.8581 -476.2192 -370.1939
#> s[-6] s[-7] s[-8] s[-9] s[-10] s[-11]
#> -303.5806 -207.384 122.0022 483.3319 620.3601 610.8525
#>
#> sigma^2: 0.009
#>
#> AIC AICc BIC
#> 1033.474 1044.807 1072.177
fit %>%
refit(as_tsibble(fdeaths)) %>%
report(reinitialise = TRUE)
#> Series: value
#> Model: ETS(M,A,A)
#> Smoothing parameters:
#> alpha = 0.0002065548
#> beta = 0.0001865257
#> gamma = 0.000118306
#>
#> Initial states:
#> l[0] b[0] s[0] s[-1] s[-2] s[-3] s[-4] s[-5]
#> 586.8764 -0.7008449 129.4235 -60.401 -108.8126 -185.465 -189.2346 -149.2135
#> s[-6] s[-7] s[-8] s[-9] s[-10] s[-11]
#> -134.8698 -70.64105 45.28081 204.0216 279.4489 240.4628
#>
#> sigma^2: 0.0118
#>
#> AIC AICc BIC
#> 903.5169 910.8854 935.3903