convertToLlamaCVFolds

For features, mean values are computed across repetitions.
For algorithms, repetitions are not supported at the moment and will result in an error.

Provides an interface to the algorithm selection benchmark library
at <http://www.aslib.net> and the 'LLAMA' package
(<https://cran.r-project.org/package=llama>) for building
algorithm selection models; see Bischl et al. (2016)
<doi:10.1016/j.artint.2016.04.003>.

Lars Kotthoff

aslib

Interface to the Algorithm Selection Benchmark Library

convertToLlamaCVFolds function

<dl><dt>asscenario</dt>
<dd>[<code>ASScenario</code>] 
Algorithm selection scenario.</dd>
<dt>measure</dt>
<dd>[<code>character(1)</code>] 
Measure to use for modelling.
Default is first measure in scenario.</dd>
<dt>feature.steps</dt>
<dd>[<code>character</code>] 
Which instance feature steps are allowed?
Default are the default instance feature steps or all steps
in case no defaults were defined.</dd>
<dt>algorithm.feature.steps</dt>
<dd>[<code>character</code>] 
Which algorithm feature steps are allowed?
Default are the default algorithm feature steps or all steps
in case no defaults were defined.</dd>
<dt>cv.splits</dt>
<dd>[<code>data.frame</code>] 
Data frame defining the split of the data into cross-validation folds,
as returned by <code>createCVSplits</code>.
Default are the splits <code>asscenario$cv.splits</code></dd></dl>

Arguments

Convert an ASScenario scenario object to a llama data object with cross-validation folds. — convertToLlamaCVFolds

<dl>

<dt>asscenario</dt>
<dd>[<code>ASScenario</code>] 
Algorithm selection scenario.</dd>


<dt>measure</dt>
<dd>[<code>character(1)</code>] 
Measure to use for modelling.
Default is first measure in scenario.</dd>


<dt>feature.steps</dt>
<dd>[<code>character</code>] 
Which instance feature steps are allowed?
Default are the default instance feature steps or all steps
in case no defaults were defined.</dd>


<dt>algorithm.feature.steps</dt>
<dd>[<code>character</code>] 
Which algorithm feature steps are allowed?
Default are the default algorithm feature steps or all steps
in case no defaults were defined.</dd>


<dt>cv.splits</dt>
<dd>[<code>data.frame</code>] 
Data frame defining the split of the data into cross-validation folds,
as returned by <code>createCVSplits</code>.
Default are the splits <code>asscenario$cv.splits</code></dd>

</dl>

convertToLlamaCVFolds: Convert an ASScenario scenario object to a llama data object with cross-validation folds.

Description

Usage

Value

Arguments