在 tidymodels 中，发送到 `recipe` 的数据与发送到 `resample` 函数的数据有何关系？

Question

在这个带有交叉验证的正则化回归的

workflow

 中，提供了两次数据：在 

recipe

 调用和 

vfold_cv

 调用中。

如果

recipe

 对数据执行一些预处理而 

vfold_cv

 函数不执行，这不会导致冲突吗？似乎 

recipe

 可以安排和扩展虚拟对象等，从而改变观察的顺序和数据框中列的数据类型。

library(tidymodels)

model <- workflow() |> 
  add_recipe(
    recipe(mpg ~ wt + cyl + drat + qsec + vs + am, data = mtcars) |> # here
      step_scale() |> 
      step_center()
  ) |> 
  add_model(
    linear_reg(penalty = tune()) |> 
      set_engine("glmnet")
  ) |> 
  tune_grid(
    resamples = vfold_cv(v = 5, data = mtcars), # and here
    #grid = 30 # doesn't capture the minimum rmse
    grid = tibble(penalty = 10^seq(-8, 1, length.out = 40))
  )
#> Warning: package 'glmnet' was built under R version 4.3.3
#> → A | warning: A correlation computation is required, but `estimate` is constant and has 0
#>                standard deviation, resulting in a divide by 0 error. `NA` will be returned.
#> There were issues with some computations   A: x1There were issues with some computations   A: x2There were issues with some computations   A: x4There were issues with some computations   A: x5

model |> 
  autoplot()

创建于 2024-04-15，使用 ^{reprex v2.0.2}

Answer 1

食谱在幕后处理这个问题

https://forum.posit.co/t/is-there-a-need-to-prep-a-recipe-when-using-vfold-cv-workflow-fit-resample-and-last-fit/ 77875/2