resample: Resample to evaluate performance of 'mpTune' chosen model

In linxihui/lazyML: Automatic machine learning algorithms and hyper-parameters selection

Description

Resample (e.g., cross validation, cross validation) to estimate the true performance of the whole mpTune model choosing precedure.

Usage

resample(obj, ...)

## S3 method for class 'mpTune.formula'
resample(obj, formula = obj$data$formula,
  data = obj$data$data, ...)

## S3 method for class 'mpTune'
resample(obj, x = obj$data$x, y = obj$data$y,
  weights = obj$data$weights, samplingFunction = createCVFolds, ...,
  returnAllSummaries = TRUE, summaryArgs = NULL, returnAllMpTunes = FALSE,
  printresampleInfo = TRUE, parallelResampling = FALSE)

Arguments

obj	A mpTune object returned from a call to mpTune
...	Further arguments to be passed to resampling
formula	A model formula. Default to obj$data$formula if any
data	Data to fit. Default to obj$data$data if any
x	Design matrix if not in obj
y	Response vector if not in obj
weights	Sample weights if not in obj
samplingFunction	Any function that returns a list of indices for training
returnAllSummaries	Should all summarizedMpTune be returned
summaryArgs	Additional argument passed to summary.mpTune
returnAllMpTunes	Should all mpTune objects be returned
printresampleInfo	If to print resampling fitting progress (not implemented yet)
parallelResampling	Should resampling fitting be parallelled (not implemented yet)

Value

A 'resampledMpTune' object with entries

• aggregatePerformancea data.frame gives an aggregated (over resamples) performance

• allResamplePerformancea data.frame gives detail performance on earch resample

• sampleIndexa list of resample indices for training

See Also

See mpTune for examples

linxihui/lazyML documentation built on May 21, 2019, 6:39 a.m.