For a recipe with at least one preprocessing operation, estimate the required parameters from a training set that can be later applied to other data sets.
prep(x, ...)
# S3 method for class 'recipe'
prep(
x,
training = NULL,
fresh = FALSE,
verbose = FALSE,
retain = TRUE,
log_changes = FALSE,
strings_as_factors = TRUE,
...
)an recipe() object.
further arguments passed to or from other methods (not currently used).
A data frame, tibble, or sparse matrix from the Matrix
package, that will be used to estimate parameters for preprocessing. See
sparse_data for more information about use of sparse data.
A logical indicating whether already trained operation should be
re-trained. If TRUE, you should pass in a data set to the argument
training.
A logical that controls whether progress is reported as operations are executed.
A logical: should the preprocessed training set be saved into
the template slot of the recipe after training? This is a good idea if
you want to add more steps later but want to avoid re-training the existing
steps. Also, it is advisable to use retain = TRUE if any steps use the
option skip = FALSE. Note that this can make the final recipe size
large. When verbose = TRUE, a message is written with the approximate
object size in memory but may be an underestimate since it does not take
environments into account.
A logical for printing a summary for each step regarding which (if any) columns were added or removed during training.
A logical: should character columns that have role
"predictor" or "outcome" be converted to factors? This affects the
preprocessed training set (when retain = TRUE) as well as the results of
bake().
A recipe whose step objects have been updated with the required quantities
(e.g. parameter estimates, model objects, etc). Also, the term_info object
is likely to be modified as the operations are executed.
Given a data set, this function estimates the required quantities and
statistics needed by any operations. prep() returns an updated recipe with
the estimates. If you are using a recipe as a preprocessor for modeling, we
highly recommend that you use a workflow() instead of manually
estimating a recipe (see the example in recipe()).
Note that missing data is handled in the steps; there is no global na.rm
option at the recipe level or in prep().
Also, if a recipe has been trained using prep() and then steps are added,
prep() will only update the new operations. If fresh = TRUE, all of the
operations will be (re)estimated.
As the steps are executed, the training set is updated. For example, if the
first step is to center the data and the second is to scale the data, the
step for scaling is given the centered data.
data(ames, package = "modeldata")
library(dplyr)
ames <- mutate(ames, Sale_Price = log10(Sale_Price))
ames_rec <-
recipe(
Sale_Price ~ Longitude + Latitude + Neighborhood + Year_Built + Central_Air,
data = ames
) %>%
step_other(Neighborhood, threshold = 0.05) %>%
step_dummy(all_nominal()) %>%
step_interact(~ starts_with("Central_Air"):Year_Built) %>%
step_ns(Longitude, Latitude, deg_free = 5)
prep(ames_rec, verbose = TRUE)
#> oper 1 step other [training]
#> oper 2 step dummy [training]
#> oper 3 step interact [training]
#> oper 4 step ns [training]
#> The retained training set is ~ 0.48 Mb in memory.
#>
#>
#> ── Recipe ──────────────────────────────────────────────────────────────────────
#>
#> ── Inputs
#> Number of variables by role
#> outcome: 1
#> predictor: 5
#>
#> ── Training information
#> Training data contained 2930 data points and no incomplete rows.
#>
#> ── Operations
#> • Collapsing factor levels for: Neighborhood | Trained
#> • Dummy variables from: Neighborhood Central_Air | Trained
#> • Interactions with: Central_Air_Y:Year_Built | Trained
#> • Natural splines on: Longitude Latitude | Trained
prep(ames_rec, log_changes = TRUE)
#> step_other (other_JgwxH): same number of columns
#>
#> step_dummy (dummy_Fe2kG):
#> new (9): Neighborhood_College_Creek, Neighborhood_Old_Town, ...
#> removed (2): Neighborhood, Central_Air
#>
#> step_interact (interact_LGGmr):
#> new (1): Central_Air_Y_x_Year_Built
#>
#> step_ns (ns_lWKqm):
#> new (10): Longitude_ns_1, Longitude_ns_2, Longitude_ns_3, ...
#> removed (2): Longitude, Latitude
#>
#>
#> ── Recipe ──────────────────────────────────────────────────────────────────────
#>
#> ── Inputs
#> Number of variables by role
#> outcome: 1
#> predictor: 5
#>
#> ── Training information
#> Training data contained 2930 data points and no incomplete rows.
#>
#> ── Operations
#> • Collapsing factor levels for: Neighborhood | Trained
#> • Dummy variables from: Neighborhood Central_Air | Trained
#> • Interactions with: Central_Air_Y:Year_Built | Trained
#> • Natural splines on: Longitude Latitude | Trained