Default formula blueprint
Source:R/blueprint-formula-default.R, R/mold.R
default_formula_blueprint.RdThis pages holds the details for the formula preprocessing blueprint. This
is the blueprint used by default from mold() if x is a formula.
Usage
default_formula_blueprint(
intercept = FALSE,
allow_novel_levels = FALSE,
indicators = "traditional",
composition = "tibble"
)
# S3 method for class 'formula'
mold(formula, data, ..., blueprint = NULL)Arguments
- intercept
A logical. Should an intercept be included in the processed data? This information is used by the
processfunction in themoldandforgefunction list.- allow_novel_levels
A logical. Should novel factor levels be allowed at prediction time? This information is used by the
cleanfunction in theforgefunction list, and is passed on toscream().- indicators
A single character string. Control how factors are expanded into dummy variable indicator columns. One of:
"traditional"- The default. Create dummy variables using the traditionalmodel.matrix()infrastructure. Generally this createsK - 1indicator columns for each factor, whereKis the number of levels in that factor."none"- Leave factor variables alone. No expansion is done."one_hot"- Create dummy variables using a one-hot encoding approach that expands unordered factors into allKindicator columns, rather thanK - 1.
- composition
Either "tibble", "matrix", or "dgCMatrix" for the format of the processed predictors. If "matrix" or "dgCMatrix" are chosen, all of the predictors must be numeric after the preprocessing method has been applied; otherwise an error is thrown.
- formula
A formula specifying the predictors and the outcomes.
- data
A data frame or matrix containing the outcomes and predictors.
- ...
Not used.
- blueprint
A preprocessing
blueprint. If left asNULL, then adefault_formula_blueprint()is used.
Details
While not different from base R, the behavior of expanding factors into
dummy variables when indicators = "traditional" and an intercept is not
present is not always intuitive and should be documented.
When an intercept is present, factors are expanded into
K-1new columns, whereKis the number of levels in the factor.When an intercept is not present, the first factor is expanded into all
Kcolumns (one-hot encoding), and the remaining factors are expanded intoK-1columns. This behavior ensures that meaningful predictions can be made for the reference level of the first factor, but is not the exact "no intercept" model that was requested. Without this behavior, predictions for the reference level of the first factor would always be forced to0when there is no intercept.
Offsets can be included in the formula method through the use of the inline
function stats::offset(). These are returned as a tibble with 1 column
named ".offset" in the $extras$offset slot of the return value.
Mold
When mold() is used with the default formula blueprint:
Predictors
The RHS of the
formulais isolated, and converted to its own 1 sided formula:~ RHS.Runs
stats::model.frame()on the RHS formula and usesdata.If
indicators = "traditional", it then runsstats::model.matrix()on the result.If
indicators = "none", factors are removed beforemodel.matrix()is run, and then added back afterwards. No interactions or inline functions involving factors are allowed.If
indicators = "one_hot", it then runsstats::model.matrix()on the result using a contrast function that creates indicator columns for all levels of all factors.If any offsets are present from using
offset(), then they are extracted withmodel_offset().If
intercept = TRUE, adds an intercept column.Coerces the result of the above steps to a tibble.
Outcomes
The LHS of the
formulais isolated, and converted to its own 1 sided formula:~ LHS.Runs
stats::model.frame()on the LHS formula and usesdata.Coerces the result of the above steps to a tibble.
Forge
When forge() is used with the default formula blueprint:
It calls
shrink()to trimnew_datato only the required columns and coercenew_datato a tibble.It calls
scream()to perform validation on the structure of the columns ofnew_data.Predictors
It runs
stats::model.frame()onnew_datausing the stored terms object corresponding to the predictors.If, in the original
mold()call,indicators = "traditional"was set, it then runsstats::model.matrix()on the result.If, in the original
mold()call,indicators = "none"was set, it runsstats::model.matrix()on the result without the factor columns, and then adds them on afterwards.If, in the original
mold()call,indicators = "one_hot"was set, it runsstats::model.matrix()on the result with a contrast function that includes indicators for all levels of all factor columns.If any offsets are present from using
offset()in the original call tomold(), then they are extracted withmodel_offset().If
intercept = TRUEin the original call tomold(), then an intercept column is added.It coerces the result of the above steps to a tibble.
Outcomes
It runs
stats::model.frame()onnew_datausing the stored terms object corresponding to the outcomes.Coerces the result to a tibble.
Differences From Base R
There are a number of differences from base R regarding how formulas are
processed by mold() that require some explanation.
Multivariate outcomes can be specified on the LHS using syntax that is
similar to the RHS (i.e. outcome_1 + outcome_2 ~ predictors).
If any complex calculations are done on the LHS and they return matrices
(like stats::poly()), then those matrices are flattened into multiple
columns of the tibble after the call to model.frame(). While this is
possible, it is not recommended, and if a large amount of preprocessing is
required on the outcomes, then you are better off
using a recipes::recipe().
Global variables are not allowed in the formula. An error will be thrown
if they are included. All terms in the formula should come from data. If
you need to use inline functions in the formula, the safest way to do so is
to prefix them with their package name, like pkg::fn(). This ensures that
the function will always be available at mold() (fit) and forge()
(prediction) time. That said, if the package is attached
(i.e. with library()), then you should be able to use the inline function
without the prefix.
By default, intercepts are not included in the predictor output from the
formula. To include an intercept, set
blueprint = default_formula_blueprint(intercept = TRUE). The rationale
for this is that many packages either always require or never allow an
intercept (for example, the earth package), and they do a large amount of
extra work to keep the user from supplying one or removing it. This
interface standardizes all of that flexibility in one place.
Examples
# ---------------------------------------------------------------------------
data("hardhat-example-data")
# ---------------------------------------------------------------------------
# Formula Example
# Call mold() with the training data
processed <- mold(
log(num_1) ~ num_2 + fac_1,
example_train,
blueprint = default_formula_blueprint(intercept = TRUE)
)
# Then, call forge() with the blueprint and the test data
# to have it preprocess the test data in the same way
forge(example_test, processed$blueprint)
#> $predictors
#> # A tibble: 2 × 4
#> `(Intercept)` num_2 fac_1b fac_1c
#> <dbl> <dbl> <dbl> <dbl>
#> 1 1 0.967 0 0
#> 2 1 0.761 0 1
#>
#> $outcomes
#> NULL
#>
#> $extras
#> $extras$offset
#> NULL
#>
#>
# Use `outcomes = TRUE` to also extract the preprocessed outcome
forge(example_test, processed$blueprint, outcomes = TRUE)
#> $predictors
#> # A tibble: 2 × 4
#> `(Intercept)` num_2 fac_1b fac_1c
#> <dbl> <dbl> <dbl> <dbl>
#> 1 1 0.967 0 0
#> 2 1 0.761 0 1
#>
#> $outcomes
#> # A tibble: 2 × 1
#> `log(num_1)`
#> <dbl>
#> 1 3.00
#> 2 3.04
#>
#> $extras
#> $extras$offset
#> NULL
#>
#>
# ---------------------------------------------------------------------------
# Factors without an intercept
# No intercept is added by default
processed <- mold(num_1 ~ fac_1 + fac_2, example_train)
# So, for factor columns, the first factor is completely expanded into all
# `K` columns (the number of levels), and the subsequent factors are expanded
# into `K - 1` columns.
processed$predictors
#> # A tibble: 12 × 4
#> fac_1a fac_1b fac_1c fac_2B
#> <dbl> <dbl> <dbl> <dbl>
#> 1 1 0 0 0
#> 2 1 0 0 1
#> 3 1 0 0 0
#> 4 1 0 0 1
#> 5 0 1 0 0
#> 6 0 1 0 1
#> 7 0 1 0 0
#> 8 0 1 0 1
#> 9 0 0 1 0
#> 10 0 0 1 1
#> 11 0 0 1 0
#> 12 0 0 1 1
# In the above example, `fac_1` is expanded into all three columns,
# `fac_2` is not. This behavior comes from `model.matrix()`, and is somewhat
# known in the R community, but can lead to a model that is difficult to
# interpret since the corresponding p-values are testing wildly different
# hypotheses.
# To get all indicators for all columns (irrespective of the intercept),
# use the `indicators = "one_hot"` option
processed <- mold(
num_1 ~ fac_1 + fac_2,
example_train,
blueprint = default_formula_blueprint(indicators = "one_hot")
)
processed$predictors
#> # A tibble: 12 × 5
#> fac_1a fac_1b fac_1c fac_2A fac_2B
#> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 1 0 0 1 0
#> 2 1 0 0 0 1
#> 3 1 0 0 1 0
#> 4 1 0 0 0 1
#> 5 0 1 0 1 0
#> 6 0 1 0 0 1
#> 7 0 1 0 1 0
#> 8 0 1 0 0 1
#> 9 0 0 1 1 0
#> 10 0 0 1 0 1
#> 11 0 0 1 1 0
#> 12 0 0 1 0 1
# It is not possible to construct a no-intercept model that expands all
# factors into `K - 1` columns using the formula method. If required, a
# recipe could be used to construct this model.
# ---------------------------------------------------------------------------
# Global variables
y <- rep(1, times = nrow(example_train))
# In base R, global variables are allowed in a model formula
frame <- model.frame(fac_1 ~ y + num_2, example_train)
head(frame)
#> fac_1 y num_2
#> 1 a 1 0.579
#> 2 a 1 0.338
#> 3 a 1 0.206
#> 4 a 1 0.546
#> 5 b 1 0.964
#> 6 b 1 0.631
# mold() does not allow them, and throws an error
try(mold(fac_1 ~ y + num_2, example_train))
#> Error in get_all_predictors(formula, data) :
#> The following predictor was not found in `data`: "y".
# ---------------------------------------------------------------------------
# Dummy variables and interactions
# By default, factor columns are expanded
# and interactions are created, both by
# calling `model.matrix()`. Some models (like
# tree based models) can take factors directly
# but still might want to use the formula method.
# In those cases, set `indicators = "none"` to not
# run `model.matrix()` on factor columns. Interactions
# are still allowed and are run on numeric columns.
bp_no_indicators <- default_formula_blueprint(indicators = "none")
processed <- mold(
~ fac_1 + num_1:num_2,
example_train,
blueprint = bp_no_indicators
)
processed$predictors
#> # A tibble: 12 × 2
#> `num_1:num_2` fac_1
#> <dbl> <fct>
#> 1 0.579 a
#> 2 0.676 a
#> 3 0.618 a
#> 4 2.18 a
#> 5 4.82 b
#> 6 3.79 b
#> 7 5.66 b
#> 8 1.66 b
#> 9 2.84 c
#> 10 0.83 c
#> 11 6.81 c
#> 12 7.42 c
# An informative error is thrown when `indicators = "none"` and
# factors are present in interaction terms or in inline functions
try(mold(num_1 ~ num_2:fac_1, example_train, blueprint = bp_no_indicators))
#> Error in mold_formula_default_process_predictors(blueprint = blueprint, :
#> Interaction terms involving factors or characters have been
#> detected on the RHS of `formula`. These are not allowed when `indicators
#> = "none"`.
#> ℹ Interactions terms involving factors were detected for "fac_1" in
#> `num_2:fac_1`.
try(mold(num_1 ~ paste0(fac_1), example_train, blueprint = bp_no_indicators))
#> Error in mold_formula_default_process_predictors(blueprint = blueprint, :
#> Functions involving factors or characters have been detected on
#> the RHS of `formula`. These are not allowed when `indicators = "none"`.
#> ℹ Functions involving factors were detected for "fac_1" in
#> `paste0(fac_1)`.
# ---------------------------------------------------------------------------
# Multivariate outcomes
# Multivariate formulas can be specified easily
processed <- mold(num_1 + log(num_2) ~ fac_1, example_train)
processed$outcomes
#> # A tibble: 12 × 2
#> num_1 `log(num_2)`
#> <int> <dbl>
#> 1 1 -0.546
#> 2 2 -1.08
#> 3 3 -1.58
#> 4 4 -0.605
#> 5 5 -0.0367
#> 6 6 -0.460
#> 7 7 -0.213
#> 8 8 -1.57
#> 9 9 -1.15
#> 10 10 -2.49
#> 11 11 -0.480
#> 12 12 -0.481
# Inline functions on the LHS are run, but any matrix
# output is flattened (like what happens in `model.matrix()`)
# (essentially this means you don't wind up with columns
# in the tibble that are matrices)
processed <- mold(poly(num_2, degree = 2) ~ fac_1, example_train)
processed$outcomes
#> # A tibble: 12 × 2
#> `poly(num_2, degree = 2).1` `poly(num_2, degree = 2).2`
#> <dbl> <dbl>
#> 1 0.0981 -0.254
#> 2 -0.177 -0.157
#> 3 -0.327 0.108
#> 4 0.0604 -0.270
#> 5 0.537 0.634
#> 6 0.157 -0.209
#> 7 0.359 0.120
#> 8 -0.325 0.103
#> 9 -0.202 -0.124
#> 10 -0.468 0.492
#> 11 0.144 -0.221
#> 12 0.143 -0.222
# TRUE
ncol(processed$outcomes) == 2
#> [1] TRUE
# Multivariate formulas specified in mold()
# carry over into forge()
forge(example_test, processed$blueprint, outcomes = TRUE)
#> $predictors
#> # A tibble: 2 × 3
#> fac_1a fac_1b fac_1c
#> <dbl> <dbl> <dbl>
#> 1 1 0 0
#> 2 0 0 1
#>
#> $outcomes
#> # A tibble: 2 × 2
#> `poly(num_2, degree = 2).1` `poly(num_2, degree = 2).2`
#> <dbl> <dbl>
#> 1 0.541 0.646
#> 2 0.306 0.00619
#>
#> $extras
#> $extras$offset
#> NULL
#>
#>
# ---------------------------------------------------------------------------
# Offsets
# Offsets are handled specially in base R, so they deserve special
# treatment here as well. You can add offsets using the inline function
# `offset()`
processed <- mold(num_1 ~ offset(num_2) + fac_1, example_train)
processed$extras$offset
#> # A tibble: 12 × 1
#> .offset
#> <dbl>
#> 1 0.579
#> 2 0.338
#> 3 0.206
#> 4 0.546
#> 5 0.964
#> 6 0.631
#> 7 0.808
#> 8 0.208
#> 9 0.316
#> 10 0.083
#> 11 0.619
#> 12 0.618
# Multiple offsets can be included, and they get added together
processed <- mold(
num_1 ~ offset(num_2) + offset(num_3),
example_train
)
identical(
processed$extras$offset$.offset,
example_train$num_2 + example_train$num_3
)
#> [1] TRUE
# Forging test data will also require
# and include the offset
forge(example_test, processed$blueprint)
#> $predictors
#> # A tibble: 2 × 0
#>
#> $outcomes
#> NULL
#>
#> $extras
#> $extras$offset
#> # A tibble: 2 × 1
#> .offset
#> <dbl>
#> 1 1.06
#> 2 0.802
#>
#>
# ---------------------------------------------------------------------------
# Intercept only
# Because `1` and `0` are intercept modifying terms, they are
# not allowed in the formula and are instead controlled by the
# `intercept` argument of the blueprint. To use an intercept
# only formula, you should supply `NULL` on the RHS of the formula.
mold(
~NULL,
example_train,
blueprint = default_formula_blueprint(intercept = TRUE)
)
#> $predictors
#> # A tibble: 12 × 1
#> `(Intercept)`
#> <dbl>
#> 1 1
#> 2 1
#> 3 1
#> 4 1
#> 5 1
#> 6 1
#> 7 1
#> 8 1
#> 9 1
#> 10 1
#> 11 1
#> 12 1
#>
#> $outcomes
#> # A tibble: 12 × 0
#>
#> $blueprint
#> Formula blueprint:
#> # Predictors: 0
#> # Outcomes: 0
#> Intercept: TRUE
#> Novel Levels: FALSE
#> Composition: tibble
#> Indicators: traditional
#>
#>
#> $extras
#> $extras$offset
#> NULL
#>
#>
# ---------------------------------------------------------------------------
# Matrix output for predictors
# You can change the `composition` of the predictor data set
bp <- default_formula_blueprint(composition = "dgCMatrix")
processed <- mold(log(num_1) ~ num_2 + fac_1, example_train, blueprint = bp)
class(processed$predictors)
#> [1] "dgCMatrix"
#> attr(,"package")
#> [1] "Matrix"