Creates a parameter set for model multiplexer tuning.
Source:R/ModelMultiplexerParamSet.R
makeModelMultiplexerParamSet.Rd
Handy way to create the param set with less typing.
The following is done automatically:
The
selected.learner
param is createdParameter names are prefixed.
The
requires
field of each param is set. This makes all parameters subordinate toselected.learner
Arguments
- multiplexer
(ModelMultiplexer)
The muliplexer learner.- ...
(ParamHelpers::ParamSet | ParamHelpers::Param)
(a) First option: Named param sets. Names must correspond to base learners. You only need to enter the parameters you want to tune without reference to theselected.learner
field in any way.
(b) Second option. Just the params you would enter in the param sets. Even shorter to create. Only works when it can be uniquely identified to which learner each of your passed parameters belongs.- .check
(logical)
Check that for each param in...
one param in found in the base learners. Default isTRUE
See also
Other multiplexer:
makeModelMultiplexer()
Other tune:
TuneControl
,
getNestedTuneResultsOptPathDf()
,
getNestedTuneResultsX()
,
getResamplingIndices()
,
getTuneResult()
,
makeModelMultiplexer()
,
makeTuneControlCMAES()
,
makeTuneControlDesign()
,
makeTuneControlGenSA()
,
makeTuneControlGrid()
,
makeTuneControlIrace()
,
makeTuneControlMBO()
,
makeTuneControlRandom()
,
makeTuneWrapper()
,
tuneParams()
,
tuneThreshold()