Generate a set of workflow objects from preprocessing and model objects

Source: R/workflow_set.R

Generate a set of workflow objects from preprocessing and model objects

workflow_set(preproc, models, cross = TRUE)

Arguments

preproc

A list (preferably named) with preprocessing objects: formulas, recipes, or workflows::workflow_variables().
models

A list (preferably named) of parsnip model specifications.
cross

A logical: should all combinations of the preprocessors and models be used to create the workflows? If FALSE, the length of preproc and models should be equal.

Value

A tibble with extra class 'workflow_set'. A new set includes four columns (but others can be added):

  • wflow_id contains character strings for the preprocessor/workflow combination. These can be changed but must be unique.

  • info is a list column with tibbles containing more specific information, including any comments added using comment_add(). This tibble also contains the workflow object (which can be easily retrieved using pull_workflow()).

  • option is a list column that will include a list of optional arguments passed to the functions from the tune package. They can be added manually via option_add() or automatically when options are passed to workflow_map().

  • result is a list column that will contain any objects produced when workflow_map() is used.

Details

The preprocessors that can be combined with the model objects can be one or more of:

Since preproc is a named list column, any combination of these can be used in that argument (i.e., preproc can be mixed types).

See also

workflow_map(), comment_add(), option_add(), as_workflow_set()

Examples

library(workflowsets)
library(workflows)
library(modeldata)
library(recipes)

#> 
#> Attaching package: ‘recipes’
#> The following object is masked from ‘package:stats’:
#> 
#>     step
library(parsnip)
library(dplyr)
library(rsample)
library(tune)
library(yardstick)

#> For binary classification, the first factor level is assumed to be the event.
#> Use the argument `event_level = "second"` to alter this as needed.

# ------------------------------------------------------------------------------

data(cells)
cells <- cells %>% dplyr::select(-case)

set.seed(1)
val_set <- validation_split(cells)

# ------------------------------------------------------------------------------

basic_recipe <-
   recipe(class ~ ., data = cells) %>%
   step_YeoJohnson(all_predictors()) %>%
   step_normalize(all_predictors())

pca_recipe <-
   basic_recipe %>%
   step_pca(all_predictors(), num_comp = tune())

ss_recipe <-
   basic_recipe %>%
   step_spatialsign(all_predictors())

# ------------------------------------------------------------------------------

knn_mod <-
   nearest_neighbor(neighbors = tune(), weight_func = tune()) %>%
   set_engine("kknn") %>%
   set_mode("classification")

lr_mod <-
   logistic_reg() %>%
   set_engine("glm")

# ------------------------------------------------------------------------------

preproc <- list(none = basic_recipe, pca = pca_recipe, sp_sign = ss_recipe)
models <- list(knn = knn_mod, logistic = lr_mod)

cell_set <- workflow_set(preproc, models, cross = TRUE)
cell_set

#> # A workflow set/tibble: 6 x 4
#>   wflow_id         info                 option    result    
#>   <chr>            <list>               <list>    <list>    
#> 1 none_knn         <tibble[,4] [1 × 4]> <opts[0]> <list [0]>
#> 2 none_logistic    <tibble[,4] [1 × 4]> <opts[0]> <list [0]>
#> 3 pca_knn          <tibble[,4] [1 × 4]> <opts[0]> <list [0]>
#> 4 pca_logistic     <tibble[,4] [1 × 4]> <opts[0]> <list [0]>
#> 5 sp_sign_knn      <tibble[,4] [1 × 4]> <opts[0]> <list [0]>
#> 6 sp_sign_logistic <tibble[,4] [1 × 4]> <opts[0]> <list [0]>

# ------------------------------------------------------------------------------
# Using variables and formulas

# Select predictors by their names
channels <- paste0("ch_", 1:4)
preproc <- purrr::map(channels, ~ workflow_variables(class, c(contains(!!.x))))
names(preproc) <- channels
preproc$everything <- class ~ .
preproc

#> $ch_1
#> $outcomes
#> <quosure>
#> expr: ^class
#> env:  0x55dc252574c8
#> 
#> $predictors
#> <quosure>
#> expr: ^c(contains("ch_1"))
#> env:  0x55dc252574c8
#> 
#> attr(,"class")
#> [1] "workflow_variables"
#> 
#> $ch_2
#> $outcomes
#> <quosure>
#> expr: ^class
#> env:  0x55dc251dfdd8
#> 
#> $predictors
#> <quosure>
#> expr: ^c(contains("ch_2"))
#> env:  0x55dc251dfdd8
#> 
#> attr(,"class")
#> [1] "workflow_variables"
#> 
#> $ch_3
#> $outcomes
#> <quosure>
#> expr: ^class
#> env:  0x55dc25169948
#> 
#> $predictors
#> <quosure>
#> expr: ^c(contains("ch_3"))
#> env:  0x55dc25169948
#> 
#> attr(,"class")
#> [1] "workflow_variables"
#> 
#> $ch_4
#> $outcomes
#> <quosure>
#> expr: ^class
#> env:  0x55dc25168320
#> 
#> $predictors
#> <quosure>
#> expr: ^c(contains("ch_4"))
#> env:  0x55dc25168320
#> 
#> attr(,"class")
#> [1] "workflow_variables"
#> 
#> $everything
#> class ~ .
#> <environment: 0x55dc27b72a78>
#> 

cell_set_by_group <- workflow_set(preproc, models["logistic"])
cell_set_by_group

#> # A workflow set/tibble: 5 x 4
#>   wflow_id            info                 option    result    
#>   <chr>               <list>               <list>    <list>    
#> 1 ch_1_logistic       <tibble[,4] [1 × 4]> <opts[0]> <list [0]>
#> 2 ch_2_logistic       <tibble[,4] [1 × 4]> <opts[0]> <list [0]>
#> 3 ch_3_logistic       <tibble[,4] [1 × 4]> <opts[0]> <list [0]>
#> 4 ch_4_logistic       <tibble[,4] [1 × 4]> <opts[0]> <list [0]>
#> 5 everything_logistic <tibble[,4] [1 × 4]> <opts[0]> <list [0]>