-
-
Notifications
You must be signed in to change notification settings - Fork 15
/
mlr_learners_classif.glmnet.Rd
245 lines (225 loc) · 11.4 KB
/
mlr_learners_classif.glmnet.Rd
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
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/LearnerClassifGlmnet.R
\name{mlr_learners_classif.glmnet}
\alias{mlr_learners_classif.glmnet}
\alias{LearnerClassifGlmnet}
\title{GLM with Elastic Net Regularization Classification Learner}
\description{
Generalized linear models with elastic net regularization.
Calls \code{\link[glmnet:glmnet]{glmnet::glmnet()}} from package \CRANpkg{glmnet}.
}
\details{
Caution: This learner is different to learners calling \code{\link[glmnet:cv.glmnet]{glmnet::cv.glmnet()}}
in that it does not use the internal optimization of parameter \code{lambda}.
Instead, \code{lambda} needs to be tuned by the user (e.g., via \CRANpkg{mlr3tuning}).
When \code{lambda} is tuned, the \code{glmnet} will be trained for each tuning iteration.
While fitting the whole path of \code{lambda}s would be more efficient, as is done
by default in \code{\link[glmnet:glmnet]{glmnet::glmnet()}}, tuning/selecting the parameter at prediction time
(using parameter \code{s}) is currently not supported in \CRANpkg{mlr3}
(at least not in efficient manner).
Tuning the \code{s} parameter is, therefore, currently discouraged.
When the data are i.i.d. and efficiency is key, we recommend using the respective
auto-tuning counterparts in \code{\link[=mlr_learners_classif.cv_glmnet]{mlr_learners_classif.cv_glmnet()}} or
\code{\link[=mlr_learners_regr.cv_glmnet]{mlr_learners_regr.cv_glmnet()}}.
However, in some situations this is not applicable, usually when data are
imbalanced or not i.i.d. (longitudinal, time-series) and tuning requires
custom resampling strategies (blocked design, stratification).
}
\section{Dictionary}{
This \link{Learner} can be instantiated via the \link[mlr3misc:Dictionary]{dictionary} \link{mlr_learners} or with the associated sugar function \code{\link[=lrn]{lrn()}}:
\if{html}{\out{<div class="sourceCode">}}\preformatted{mlr_learners$get("classif.glmnet")
lrn("classif.glmnet")
}\if{html}{\out{</div>}}
}
\section{Meta Information}{
\itemize{
\item Task type: \dQuote{classif}
\item Predict Types: \dQuote{response}, \dQuote{prob}
\item Feature Types: \dQuote{logical}, \dQuote{integer}, \dQuote{numeric}
\item Required Packages: \CRANpkg{mlr3}, \CRANpkg{mlr3learners}, \CRANpkg{glmnet}
}
}
\section{Parameters}{
\tabular{lllll}{
Id \tab Type \tab Default \tab Levels \tab Range \cr
alpha \tab numeric \tab 1 \tab \tab \eqn{[0, 1]}{[0, 1]} \cr
big \tab numeric \tab 9.9e+35 \tab \tab \eqn{(-\infty, \infty)}{(-Inf, Inf)} \cr
devmax \tab numeric \tab 0.999 \tab \tab \eqn{[0, 1]}{[0, 1]} \cr
dfmax \tab integer \tab - \tab \tab \eqn{[0, \infty)}{[0, Inf)} \cr
eps \tab numeric \tab 1e-06 \tab \tab \eqn{[0, 1]}{[0, 1]} \cr
epsnr \tab numeric \tab 1e-08 \tab \tab \eqn{[0, 1]}{[0, 1]} \cr
exact \tab logical \tab FALSE \tab TRUE, FALSE \tab - \cr
exclude \tab integer \tab - \tab \tab \eqn{[1, \infty)}{[1, Inf)} \cr
exmx \tab numeric \tab 250 \tab \tab \eqn{(-\infty, \infty)}{(-Inf, Inf)} \cr
fdev \tab numeric \tab 1e-05 \tab \tab \eqn{[0, 1]}{[0, 1]} \cr
gamma \tab numeric \tab 1 \tab \tab \eqn{(-\infty, \infty)}{(-Inf, Inf)} \cr
intercept \tab logical \tab TRUE \tab TRUE, FALSE \tab - \cr
lambda \tab untyped \tab - \tab \tab - \cr
lambda.min.ratio \tab numeric \tab - \tab \tab \eqn{[0, 1]}{[0, 1]} \cr
lower.limits \tab untyped \tab - \tab \tab - \cr
maxit \tab integer \tab 100000 \tab \tab \eqn{[1, \infty)}{[1, Inf)} \cr
mnlam \tab integer \tab 5 \tab \tab \eqn{[1, \infty)}{[1, Inf)} \cr
mxit \tab integer \tab 100 \tab \tab \eqn{[1, \infty)}{[1, Inf)} \cr
mxitnr \tab integer \tab 25 \tab \tab \eqn{[1, \infty)}{[1, Inf)} \cr
nlambda \tab integer \tab 100 \tab \tab \eqn{[1, \infty)}{[1, Inf)} \cr
newoffset \tab untyped \tab - \tab \tab - \cr
offset \tab untyped \tab NULL \tab \tab - \cr
penalty.factor \tab untyped \tab - \tab \tab - \cr
pmax \tab integer \tab - \tab \tab \eqn{[0, \infty)}{[0, Inf)} \cr
pmin \tab numeric \tab 1e-09 \tab \tab \eqn{[0, 1]}{[0, 1]} \cr
prec \tab numeric \tab 1e-10 \tab \tab \eqn{(-\infty, \infty)}{(-Inf, Inf)} \cr
relax \tab logical \tab FALSE \tab TRUE, FALSE \tab - \cr
s \tab numeric \tab 0.01 \tab \tab \eqn{[0, \infty)}{[0, Inf)} \cr
standardize \tab logical \tab TRUE \tab TRUE, FALSE \tab - \cr
standardize.response \tab logical \tab FALSE \tab TRUE, FALSE \tab - \cr
thresh \tab numeric \tab 1e-07 \tab \tab \eqn{[0, \infty)}{[0, Inf)} \cr
trace.it \tab integer \tab 0 \tab \tab \eqn{[0, 1]}{[0, 1]} \cr
type.gaussian \tab character \tab - \tab covariance, naive \tab - \cr
type.logistic \tab character \tab - \tab Newton, modified.Newton \tab - \cr
type.multinomial \tab character \tab - \tab ungrouped, grouped \tab - \cr
upper.limits \tab untyped \tab - \tab \tab - \cr
}
}
\section{Internal Encoding}{
Starting with \CRANpkg{mlr3} v0.5.0, the order of class labels is reversed prior to
model fitting to comply to the \code{\link[stats:glm]{stats::glm()}} convention that the negative class is provided
as the first factor level.
}
\examples{
if (requireNamespace("glmnet", quietly = TRUE)) {
# Define the Learner and set parameter values
learner = lrn("classif.glmnet")
print(learner)
# Define a Task
task = tsk("sonar")
# Create train and test set
ids = partition(task)
# Train the learner on the training ids
learner$train(task, row_ids = ids$train)
# print the model
print(learner$model)
# importance method
if("importance" \%in\% learner$properties) print(learner$importance)
# Make predictions for the test rows
predictions = learner$predict(task, row_ids = ids$test)
# Score the predictions
predictions$score()
}
}
\references{
Friedman J, Hastie T, Tibshirani R (2010).
\dQuote{Regularization Paths for Generalized Linear Models via Coordinate Descent.}
\emph{Journal of Statistical Software}, \bold{33}(1), 1--22.
\doi{10.18637/jss.v033.i01}.
}
\seealso{
\itemize{
\item Chapter in the \href{https://mlr3book.mlr-org.com/}{mlr3book}:
\url{https://mlr3book.mlr-org.com/chapters/chapter2/data_and_basic_modeling.html#sec-learners}
\item Package \href{https://github.com/mlr-org/mlr3extralearners}{mlr3extralearners} for more learners.
\item \link[mlr3misc:Dictionary]{Dictionary} of \link[=Learner]{Learners}: \link{mlr_learners}
\item \code{as.data.table(mlr_learners)} for a table of available \link[=Learner]{Learners} in the running session (depending on the loaded packages).
\item \CRANpkg{mlr3pipelines} to combine learners with pre- and postprocessing steps.
\item Extension packages for additional task types:
\itemize{
\item \CRANpkg{mlr3proba} for probabilistic supervised regression and survival analysis.
\item \CRANpkg{mlr3cluster} for unsupervised clustering.
}
\item \CRANpkg{mlr3tuning} for tuning of hyperparameters, \CRANpkg{mlr3tuningspaces}
for established default tuning spaces.
}
Other Learner:
\code{\link{mlr_learners_classif.cv_glmnet}},
\code{\link{mlr_learners_classif.kknn}},
\code{\link{mlr_learners_classif.lda}},
\code{\link{mlr_learners_classif.log_reg}},
\code{\link{mlr_learners_classif.multinom}},
\code{\link{mlr_learners_classif.naive_bayes}},
\code{\link{mlr_learners_classif.nnet}},
\code{\link{mlr_learners_classif.qda}},
\code{\link{mlr_learners_classif.ranger}},
\code{\link{mlr_learners_classif.svm}},
\code{\link{mlr_learners_classif.xgboost}},
\code{\link{mlr_learners_regr.cv_glmnet}},
\code{\link{mlr_learners_regr.glmnet}},
\code{\link{mlr_learners_regr.kknn}},
\code{\link{mlr_learners_regr.km}},
\code{\link{mlr_learners_regr.lm}},
\code{\link{mlr_learners_regr.nnet}},
\code{\link{mlr_learners_regr.ranger}},
\code{\link{mlr_learners_regr.svm}},
\code{\link{mlr_learners_regr.xgboost}}
}
\concept{Learner}
\section{Super classes}{
\code{\link[mlr3:Learner]{mlr3::Learner}} -> \code{\link[mlr3:LearnerClassif]{mlr3::LearnerClassif}} -> \code{LearnerClassifGlmnet}
}
\section{Methods}{
\subsection{Public methods}{
\itemize{
\item \href{#method-LearnerClassifGlmnet-new}{\code{LearnerClassifGlmnet$new()}}
\item \href{#method-LearnerClassifGlmnet-selected_features}{\code{LearnerClassifGlmnet$selected_features()}}
\item \href{#method-LearnerClassifGlmnet-clone}{\code{LearnerClassifGlmnet$clone()}}
}
}
\if{html}{\out{
<details><summary>Inherited methods</summary>
<ul>
<li><span class="pkg-link" data-pkg="mlr3" data-topic="Learner" data-id="base_learner"><a href='../../mlr3/html/Learner.html#method-Learner-base_learner'><code>mlr3::Learner$base_learner()</code></a></span></li>
<li><span class="pkg-link" data-pkg="mlr3" data-topic="Learner" data-id="format"><a href='../../mlr3/html/Learner.html#method-Learner-format'><code>mlr3::Learner$format()</code></a></span></li>
<li><span class="pkg-link" data-pkg="mlr3" data-topic="Learner" data-id="help"><a href='../../mlr3/html/Learner.html#method-Learner-help'><code>mlr3::Learner$help()</code></a></span></li>
<li><span class="pkg-link" data-pkg="mlr3" data-topic="Learner" data-id="predict"><a href='../../mlr3/html/Learner.html#method-Learner-predict'><code>mlr3::Learner$predict()</code></a></span></li>
<li><span class="pkg-link" data-pkg="mlr3" data-topic="Learner" data-id="predict_newdata"><a href='../../mlr3/html/Learner.html#method-Learner-predict_newdata'><code>mlr3::Learner$predict_newdata()</code></a></span></li>
<li><span class="pkg-link" data-pkg="mlr3" data-topic="Learner" data-id="print"><a href='../../mlr3/html/Learner.html#method-Learner-print'><code>mlr3::Learner$print()</code></a></span></li>
<li><span class="pkg-link" data-pkg="mlr3" data-topic="Learner" data-id="reset"><a href='../../mlr3/html/Learner.html#method-Learner-reset'><code>mlr3::Learner$reset()</code></a></span></li>
<li><span class="pkg-link" data-pkg="mlr3" data-topic="Learner" data-id="train"><a href='../../mlr3/html/Learner.html#method-Learner-train'><code>mlr3::Learner$train()</code></a></span></li>
</ul>
</details>
}}
\if{html}{\out{<hr>}}
\if{html}{\out{<a id="method-LearnerClassifGlmnet-new"></a>}}
\if{latex}{\out{\hypertarget{method-LearnerClassifGlmnet-new}{}}}
\subsection{Method \code{new()}}{
Creates a new instance of this \link[R6:R6Class]{R6} class.
\subsection{Usage}{
\if{html}{\out{<div class="r">}}\preformatted{LearnerClassifGlmnet$new()}\if{html}{\out{</div>}}
}
}
\if{html}{\out{<hr>}}
\if{html}{\out{<a id="method-LearnerClassifGlmnet-selected_features"></a>}}
\if{latex}{\out{\hypertarget{method-LearnerClassifGlmnet-selected_features}{}}}
\subsection{Method \code{selected_features()}}{
Returns the set of selected features as reported by \code{\link[glmnet:predict.glmnet]{glmnet::predict.glmnet()}}
with \code{type} set to \code{"nonzero"}.
\subsection{Usage}{
\if{html}{\out{<div class="r">}}\preformatted{LearnerClassifGlmnet$selected_features(lambda = NULL)}\if{html}{\out{</div>}}
}
\subsection{Arguments}{
\if{html}{\out{<div class="arguments">}}
\describe{
\item{\code{lambda}}{(\code{numeric(1)})\cr
Custom \code{lambda}, defaults to the active lambda depending on parameter set.}
}
\if{html}{\out{</div>}}
}
\subsection{Returns}{
(\code{character()}) of feature names.
}
}
\if{html}{\out{<hr>}}
\if{html}{\out{<a id="method-LearnerClassifGlmnet-clone"></a>}}
\if{latex}{\out{\hypertarget{method-LearnerClassifGlmnet-clone}{}}}
\subsection{Method \code{clone()}}{
The objects of this class are cloneable with this method.
\subsection{Usage}{
\if{html}{\out{<div class="r">}}\preformatted{LearnerClassifGlmnet$clone(deep = FALSE)}\if{html}{\out{</div>}}
}
\subsection{Arguments}{
\if{html}{\out{<div class="arguments">}}
\describe{
\item{\code{deep}}{Whether to make a deep clone.}
}
\if{html}{\out{</div>}}
}
}
}