get_learners_internal: Internal function to get the learners

In hrvg/RiverML: Machine Learning for River Science

Description

The following learners "classif.h2o.glm", "classif.lda", "classif.mda", "classif.naiveBayes", "classif.IBk", "classif.kknn", "classif.featureless" do not get transformed into a TuneWrapper. The following learners c("classif.h2o.gbm", "classif.h2o.deeplearning", "classif.nnTrain", "classif.xgboost") are tuned with mlr::makeTuneControlRandom(). All other learners are tuned with mlr::makeTuneControlGrid(). Setting the possible hyper-parameters values is handled by get_ps()

Usage

get_learners_internal(
  lrn_ids,
  data,
  inner_resampling,
  grid_resolution,
  .info = FALSE,
  pca = FALSE,
  corr = FALSE,
  prob = TRUE,
  randomit = 100,
  mes
)

Arguments

lrn_ids	character, list of mlr learner idenfitiers
data	a named list with two elements data and labels
inner_resampling	resampleDesc from mlr, the inner folds of the nested resampling
grid_resolution	numeric, defines the granularity of the discrete tuning grid
.info	logical, controls the amount of information printed when tuning
pca	logical, is a PCA performed
corr	logical, are highly correlated predictors removed
prob	logical, controls the type of output, if TRUE probabilities, if FALSE response
randomit	numeric, the number of iteration for the random discrete tuning
mes	mlr list of measure to compute while tuning, the learner are tuning against the first element

Value

a list of mlr learners

hrvg/RiverML documentation built on Oct. 12, 2020, 10:40 a.m.