mlr_learners_graph

GraphLearner

A <code><a rd-options="mlr3:Learner" href="/link/Learner?package=mlr3pipelines&version=0.3.0&to=mlr3%3ALearner" data-mini-rdoc="mlr3:Learner::Learner">Learner</a></code> that encapsulates a <code><a rd-options="" href="/link/Graph?package=mlr3pipelines&version=0.3.0" data-mini-rdoc="mlr3pipelines::Graph">Graph</a></code> to be used in
<a rd-options="mlr3:mlr3-package" href="/link/mlr3?package=mlr3pipelines&version=0.3.0&to=mlr3%3Amlr3-package" data-mini-rdoc="mlr3:mlr3-package::mlr3">mlr3</a> resampling and benchmarks.
The Graph must return a single <code><a rd-options="mlr3:Prediction" href="/link/Prediction?package=mlr3pipelines&version=0.3.0&to=mlr3%3APrediction" data-mini-rdoc="mlr3:Prediction::Prediction">Prediction</a></code> on its <code>$predict()</code>
call. The result of the <code>$train()</code> call is discarded, only the
internal state changes during training are used.
The <code>predict_type</code> of a <code><a rd-options="" href="/link/GraphLearner?package=mlr3pipelines&version=0.3.0" data-mini-rdoc="mlr3pipelines::GraphLearner">GraphLearner</a></code> can be obtained or set via it's <code>predict_type</code> active binding.
Setting a new predict type will try to set the <code>predict_type</code> in all relevant
<code><a rd-options="" href="/link/PipeOp?package=mlr3pipelines&version=0.3.0" data-mini-rdoc="mlr3pipelines::PipeOp">PipeOp</a></code> / <code><a rd-options="mlr3:Learner" href="/link/Learner?package=mlr3pipelines&version=0.3.0&to=mlr3%3ALearner" data-mini-rdoc="mlr3:Learner::Learner">Learner</a></code> encapsulated within the <code><a rd-options="" href="/link/Graph?package=mlr3pipelines&version=0.3.0" data-mini-rdoc="mlr3pipelines::Graph">Graph</a></code>.
Similarly, the predict_type of a Graph will always be the smallest denominator in the <code><a rd-options="" href="/link/Graph?package=mlr3pipelines&version=0.3.0" data-mini-rdoc="mlr3pipelines::Graph">Graph</a></code>.

Dataflow programming toolkit that enriches 'mlr3' with a diverse
set of pipelining operators ('PipeOps') that can be composed into graphs.
Operations exist for data preprocessing, model fitting, and ensemble
learning. Graphs can themselves be treated as 'mlr3' 'Learners' and can
therefore be resampled, benchmarked, and tuned.

Martin Binder

mlr3pipelines

Preprocessing Operators and Pipelines for 'mlr3'

Florian Pfisterer

Lennart Schneider

Bernd Bischl

Michel Lang

Susanne Dandl

mlr_learners_graph function

<code><a rd-options="" href="/link/R6Class?package=mlr3pipelines&version=0.3.0" data-mini-rdoc="mlr3pipelines::R6Class">R6Class</a></code> object inheriting from <code><a rd-options="mlr3:Learner" href="/link/mlr3%3A%3ALearner?package=mlr3pipelines&version=0.3.0&to=mlr3%3ALearner" data-mini-rdoc="mlr3:Learner::mlr3::Learner">mlr3::Learner</a></code>.

Format

A <code><a rd-options='mlr3:Learner' href='Learner'>Learner</a></code> that encapsulates a <code><a rd-options='' href='Graph'>Graph</a></code> to be used in
<a rd-options='mlr3:mlr3-package' href='mlr3'>mlr3</a> resampling and benchmarks.
The Graph must return a single <code><a rd-options='mlr3:Prediction' href='Prediction'>Prediction</a></code> on its <code>$predict()</code>
call. The result of the <code>$train()</code> call is discarded, only the
internal state changes during training are used.
The <code>predict_type</code> of a <code><a rd-options='' href='GraphLearner'>GraphLearner</a></code> can be obtained or set via it's <code>predict_type</code> active binding.
Setting a new predict type will try to set the <code>predict_type</code> in all relevant
<code><a rd-options='' href='PipeOp'>PipeOp</a></code> / <code><a rd-options='mlr3:Learner' href='Learner'>Learner</a></code> encapsulated within the <code><a rd-options='' href='Graph'>Graph</a></code>.
Similarly, the predict_type of a Graph will always be the smallest denominator in the <code><a rd-options='' href='Graph'>Graph</a></code>.

GraphLearner — mlr_learners_graph

<code><a rd-options='' href='R6Class'>R6Class</a></code> object inheriting from <code><a rd-options='mlr3:Learner' href='mlr3::Learner'>mlr3::Learner</a></code>.

mlr_learners_graph: GraphLearner

Description

Arguments

Format

See Also