kf_xval: k-fold cross validation

In nunompmoniz/autoresampling: Automated Resampling Strategies for Imbalanced Learning

Description

kf_xval applies a k-fold cross-validation methodology to a given data set.

Usage

kf_xval(
  x,
  form,
  nfolds,
  FUN,
  shuffle.rows = TRUE,
  average_results = FALSE,
  seedlock = FALSE,
  percTest = NULL,
  ...
)

Arguments

x	data: embedded time series
form	A model formula
nfolds	no of folds
FUN	function to apply to each iteration's train and validation. Typically FUN is a workflow where a predictive model is applied in a training set and the model is evaluated in the validation set.
shuffle.rows	Boolean for shuffling rows in cross-validation methodology. Default TRUE
average_results	Boolean for return average results of metrics. Default FALSE
seedlock	Random seed. Defaults to FALSE (appointing the seed 123)
percTest	Percentage for holdout evaluation (generalisation). Default is NULL
...	further parameters to FUN

nunompmoniz/autoresampling documentation built on April 26, 2021, 4:43 a.m.