accuracy by horizon

[robjhyndman]

MRE:

library(fpp3)
#> ── Attaching packages ──────────────────────────────────────────── fpp3 0.4.0 ──
#> ✓ tibble      3.1.6          ✓ tsibble     1.1.1     
#> ✓ dplyr       1.0.7          ✓ tsibbledata 0.4.0.9000
#> ✓ tidyr       1.1.4          ✓ feasts      0.2.2     
#> ✓ lubridate   1.8.0          ✓ fable       0.3.1     
#> ✓ ggplot2     3.3.5
#> ── Conflicts ───────────────────────────────────────────────── fpp3_conflicts ──
#> x lubridate::date()    masks base::date()
#> x dplyr::filter()      masks stats::filter()
#> x tsibble::intersect() masks base::intersect()
#> x tsibble::interval()  masks lubridate::interval()
#> x dplyr::lag()         masks stats::lag()
#> x tsibble::setdiff()   masks base::setdiff()
#> x tsibble::union()     masks base::union()

fit <- aus_arrivals %>%
  filter(year(Quarter) <= 2009) %>% 
  model(ets = ETS(Arrivals))
fc <- fit %>% 
  forecast(h = 11) %>%
  group_by(Origin, .model) %>%
  mutate(h = row_number()) %>%
  ungroup()
fc %>% 
  accuracy(aus_arrivals, by=c(".model","h"))
#> Error: The result is not a valid tsibble.
#> ℹ Do you need `as_tibble()` to work with data frame?

^{Created on 2022-01-12 by the reprex package (v2.0.1)}

Not a problem if there is no key in the fable other than .model.

[mitchelloharawild]

Added. Previously by was assumed to only be for adding new grouping structures, like h, and required all existing keys to be used. This is no longer required.

library(fpp3)
#> ── Attaching packages ──────────────────────────────────────────── fpp3 0.4.0 ──
#> ✓ tibble      3.1.6          ✓ tsibble     1.0.1     
#> ✓ dplyr       1.0.7          ✓ tsibbledata 0.4.0.9000
#> ✓ tidyr       1.1.4          ✓ feasts      0.2.2.9000
#> ✓ lubridate   1.8.0          ✓ fable       0.3.1     
#> ✓ ggplot2     3.3.5.9000
#> ── Conflicts ───────────────────────────────────────────────── fpp3_conflicts ──
#> x lubridate::date()    masks base::date()
#> x dplyr::filter()      masks stats::filter()
#> x tsibble::intersect() masks base::intersect()
#> x tsibble::interval()  masks lubridate::interval()
#> x dplyr::lag()         masks stats::lag()
#> x tsibble::setdiff()   masks base::setdiff()
#> x tsibble::union()     masks base::union()
fit <- aus_arrivals %>%
  filter(year(Quarter) <= 2009) %>% 
  model(ets = ETS(Arrivals))
fc <- fit %>% 
  forecast(h = 11) %>%
  group_by(Origin, .model) %>%
  mutate(h = row_number()) %>%
  ungroup()
fc %>% 
  accuracy(aus_arrivals, by=c(".model","h"))
#> # A tibble: 11 × 11
#>    .model     h .type     ME    RMSE     MAE    MPE  MAPE  MASE RMSSE  ACF1
#>    <chr>  <int> <chr>  <dbl>   <dbl>   <dbl>  <dbl> <dbl> <dbl> <dbl> <dbl>
#>  1 ets        1 Test   6570.  77733.  61636.  -6.26  40.6  4.17  3.89 0.393
#>  2 ets        2 Test  -3355. 120400.  88372. -32.5   73.0  6.02  6.05 0.129
#>  3 ets        3 Test   8025. 136534.  94424. -18.2   58.8  6.47  6.88 0.177
#>  4 ets        4 Test   9444. 127585. 102493. -16.9   62.5  6.96  6.40 0.320
#>  5 ets        5 Test   1983.  88967.  73629. -12.6   52.9  5.03  4.48 0.379
#>  6 ets        6 Test  -1959. 131268.  97567. -45.0   91.8  6.66  6.62 0.138
#>  7 ets        7 Test  -1575. 141055. 103902. -27.6   74.6  7.08  7.12 0.235
#>  8 ets        8 Test   6508. 139005. 113831. -21.8   74.1  7.72  7.01 0.312
#>  9 ets        9 Test  11618.  99632.  81964.  -5.57  55.8  5.58  5.04 0.429
#> 10 ets       10 Test    387. 139894. 106160. -40.6   96.4  7.26  7.11 0.135
#> 11 ets       11 Test   4697. 150610. 112844. -22.8   79.9  7.75  7.68 0.255

^{Created on 2022-01-12 by the reprex package (v2.0.0)}