1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
|
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/selections.R
\name{selections}
\alias{selections}
\alias{selection}
\title{Methods for Selecting Variables in Step Functions}
\description{
When selecting variables or model terms in \code{step}
functions, \code{dplyr}-like tools are used. The \emph{selector} functions
can choose variables based on their name, current role, data
type, or any combination of these. The selectors are passed as
any other argument to the step. If the variables are explicitly
stated in the step function, this might be similar to:
\preformatted{
recipe( ~ ., data = USArrests) \%>\%
step_pca(Murder, Assault, UrbanPop, Rape, num_comp = 3)
}
The first four arguments indicate which variables should be
used in the PCA while the last argument is a specific argument
to \code{\link[=step_pca]{step_pca()}}.
Note that:
\enumerate{
\item These arguments are not evaluated until the \code{prep}
function for the step is executed.
\item The \code{dplyr}-like syntax allows for negative signs to
exclude variables (e.g. \code{-Murder}) and the set of selectors will
processed in order.
\item A leading exclusion in these arguments (e.g. \code{-Murder})
has the effect of adding all variables to the list except the
excluded variable(s).
}
Also, select helpers from the \code{tidyselect} package can also be used:
\code{\link[tidyselect:starts_with]{tidyselect::starts_with()}}, \code{\link[tidyselect:starts_with]{tidyselect::ends_with()}},
\code{\link[tidyselect:starts_with]{tidyselect::contains()}}, \code{\link[tidyselect:starts_with]{tidyselect::matches()}},
\code{\link[tidyselect:starts_with]{tidyselect::num_range()}}, \code{\link[tidyselect:everything]{tidyselect::everything()}},
\code{\link[tidyselect:one_of]{tidyselect::one_of()}}, \code{\link[tidyselect:all_of]{tidyselect::all_of()}}, and
\code{\link[tidyselect:all_of]{tidyselect::any_of()}}
For example:
\preformatted{
recipe(Species ~ ., data = iris) \%>\%
step_center(starts_with("Sepal"), -contains("Width"))
}
would only select \code{Sepal.Length}
Columns of the design matrix that may not exist when the step
is coded can also be selected. For example, when using
\code{step_pca()}, the number of columns created by feature extraction
may not be known when subsequent steps are defined. In this
case, using \code{matches("^PC")} will select all of the columns
whose names start with "PC" \emph{once those columns are created}.
There are sets of recipes specific functions that can be used to select
variables based on their role or type: \code{\link[=has_role]{has_role()}} and
\code{\link[=has_type]{has_type()}}. For convenience, there are also functions that are
more specific: \code{\link[=all_numeric]{all_numeric()}}, \code{\link[=all_nominal]{all_nominal()}},
\code{\link[=all_predictors]{all_predictors()}}, and \code{\link[=all_outcomes]{all_outcomes()}}. These can be used in
conjunction with the previous functions described for selecting
variables using their names:
\preformatted{
data(biomass)
recipe(HHV ~ ., data = biomass) \%>\%
step_center(all_numeric(), -all_outcomes())
}
This results in all the numeric predictors: carbon, hydrogen,
oxygen, nitrogen, and sulfur.
If a role for a variable has not been defined, it will never be
selected using role-specific selectors.
Selectors can be used in \code{\link[=step_interact]{step_interact()}} in similar ways but
must be embedded in a model formula (as opposed to a sequence
of selectors). For example, the interaction specification
could be \code{~ starts_with("Species"):Sepal.Width}. This can be
useful if \code{Species} was converted to dummy variables
previously using \code{\link[=step_dummy]{step_dummy()}}. The implementation of
\code{step_interact()} is special, and is more restricted than
the other step functions. Only the selector functions from
recipes and tidyselect are allowed. User defined selector functions
will not be recognized. Additionally, the tidyselect domain specific
language is not recognized here, meaning that \code{&}, \code{|}, \code{!}, and \code{-}
will not work.
}
|
