0

我有一个包含几个模型的 mable：

pacman::p_load(tidyverse, fable, fpp3, feasts)
retail_data |> autoplot() + facet_wrap(vars(State))
retail_train <- retail_data |> filter(year(Month) <= max(year(Month)) - 4)
ft.mod <- retail_train |> model(
  drift = RW(Turnover ~ drift())
)

ft.mod
# A mable: 8 x 3
# Key:     State, Industry [8]
  State                        Industry                       drift
  <chr>                        <chr>                        <model>
1 Australian Capital Territory Takeaway food services <RW w/ drift>
2 New South Wales              Takeaway food services <RW w/ drift>
3 Northern Territory           Takeaway food services <RW w/ drift>
4 Queensland                   Takeaway food services <RW w/ drift>
5 South Australia              Takeaway food services <RW w/ drift>
6 Tasmania                     Takeaway food services <RW w/ drift>
7 Victoria                     Takeaway food services <RW w/ drift>
8 Western Australia            Takeaway food services <RW w/ drift>

我可以像这样可视化单个模型的残差：

ft.mod |> filter(State == 'Victoria') |> gg_tsresiduals() # plot appears in pane

但是，如果我尝试做一个：

ft.mod |> gg_tsresiduals()
Error: gg_tsresiduals() must be used with a mable containing only one model.

我尝试使用map()：

ft.mod |> map( ~ gg_tsresiduals(.x))
Error: gg_tsresiduals() must be used with a mable containing only one model.

如何使用 map 迭代每个模型gg_tsresiduals()？

4

0 回答 0