rap with models

rap-demo.R

suppressPackageStartupMessages({
  library(rap)
  library(AmesHousing)
  library(rsample)
  library(parsnip)
  library(dplyr)
})

ames <- make_ames()

set.seed(4595)
data_split <- initial_split(ames, strata = "Sale_Price", prop = .75)

ames_train <- training(data_split)
ames_test  <- testing(data_split)

spec_lin_reg <- linear_reg()
spec_lm <- set_engine(spec_lin_reg, "lm")

ames_train_log <- ames_train %>%
  mutate(Sale_Price_Log = log10(Sale_Price))

set.seed(2453)
cv_splits <- vfold_cv(
  data = ames_train_log, 
  v = 10, 
  strata = "Sale_Price"
)

# for each split, fit a model
cv_splits
#> #  10-fold cross-validation using stratification 
#> # A tibble: 10 x 2
#>    splits           id    
#>    <list>           <chr> 
#>  1 <split [2K/222]> Fold01
#>  2 <split [2K/222]> Fold02
#>  3 <split [2K/222]> Fold03
#>  4 <split [2K/222]> Fold04
#>  5 <split [2K/222]> Fold05
#>  6 <split [2K/219]> Fold06
#>  7 <split [2K/219]> Fold07
#>  8 <split [2K/217]> Fold08
#>  9 <split [2K/217]> Fold09
#> 10 <split [2K/217]> Fold10

cv_splits %>%
  rap(
    models = ~fit(spec_lm, Sale_Price_Log ~ Latitude + Longitude, analysis(splits)),
    assess = ~assessment(splits),
    pred   = ~predict(models, new_data = assess),
    rmse   = double() ~ rmse_vec(assess$Sale_Price_Log, pred$.pred)
  )
#> #  10-fold cross-validation using stratification 
#> # A tibble: 10 x 6
#>    splits          id     models   assess            pred              rmse
#>    <list>          <chr>  <list>   <list>            <list>           <dbl>
#>  1 <split [2K/222… Fold01 <fit[+]> <tibble [222 × 8… <tibble [222 × … 0.150
#>  2 <split [2K/222… Fold02 <fit[+]> <tibble [222 × 8… <tibble [222 × … 0.164
#>  3 <split [2K/222… Fold03 <fit[+]> <tibble [222 × 8… <tibble [222 × … 0.160
#>  4 <split [2K/222… Fold04 <fit[+]> <tibble [222 × 8… <tibble [222 × … 0.158
#>  5 <split [2K/222… Fold05 <fit[+]> <tibble [222 × 8… <tibble [222 × … 0.161
#>  6 <split [2K/219… Fold06 <fit[+]> <tibble [219 × 8… <tibble [219 × … 0.156
#>  7 <split [2K/219… Fold07 <fit[+]> <tibble [219 × 8… <tibble [219 × … 0.179
#>  8 <split [2K/217… Fold08 <fit[+]> <tibble [217 × 8… <tibble [217 × … 0.159
#>  9 <split [2K/217… Fold09 <fit[+]> <tibble [217 × 8… <tibble [217 × … 0.168
#> 10 <split [2K/217… Fold10 <fit[+]> <tibble [217 × 8… <tibble [217 × … 0.158

# compare against
cv_splits %>%
  mutate(
    models = map(splits, ~fit(spec_lm, Sale_Price_Log ~ Latitude + Longitude, analysis(.x))),
    assess = map(splits, assessment),
    pred   = map2(models, assess, ~predict(.x, .y)),
    rmse   = map2_dbl(assess, pred, ~rmse_vec(.x$Sale_Price_Log, .y$.pred))
  )
#> #  10-fold cross-validation using stratification 
#> # A tibble: 10 x 6
#>    splits          id     models   assess            pred              rmse
#>  * <list>          <chr>  <list>   <list>            <list>           <dbl>
#>  1 <split [2K/222… Fold01 <fit[+]> <tibble [222 × 8… <tibble [222 × … 0.150
#>  2 <split [2K/222… Fold02 <fit[+]> <tibble [222 × 8… <tibble [222 × … 0.164
#>  3 <split [2K/222… Fold03 <fit[+]> <tibble [222 × 8… <tibble [222 × … 0.160
#>  4 <split [2K/222… Fold04 <fit[+]> <tibble [222 × 8… <tibble [222 × … 0.158
#>  5 <split [2K/222… Fold05 <fit[+]> <tibble [222 × 8… <tibble [222 × … 0.161
#>  6 <split [2K/219… Fold06 <fit[+]> <tibble [219 × 8… <tibble [219 × … 0.156
#>  7 <split [2K/219… Fold07 <fit[+]> <tibble [219 × 8… <tibble [219 × … 0.179
#>  8 <split [2K/217… Fold08 <fit[+]> <tibble [217 × 8… <tibble [217 × … 0.159
#>  9 <split [2K/217… Fold09 <fit[+]> <tibble [217 × 8… <tibble [217 × … 0.168
#> 10 <split [2K/217… Fold10 <fit[+]> <tibble [217 × 8… <tibble [217 × … 0.158