ensemble_dnnet: An ensemble model of DNNs

In SkadiEye/deepTL: Deep Treatment Learning

An ensemble model of DNNs

Description

Fit a an ensemble model of DNNs

Usage

ensemble_dnnet(
  object,
  n.ensemble = 100,
  esCtrl,
  random.nbatch = FALSE,
  random.nbatch.limit = NULL,
  unbalance.trt = c("None", "Over", "Under", "Balance")[1],
  bootstrap = TRUE,
  prop.train = 1,
  prop.keep = 1,
  best.opti = TRUE,
  min.keep = 10,
  verbose = TRUE
)

Arguments

object	A dnnetInput or a dnnetSurvInput object, the training set.
n.ensemble	The number of DNNs to be fit
esCtrl	A list of parameters to be passed to dnnet.
random.nbatch	An indicator whether n.batch is randomly selected.
random.nbatch.limit	Minimum and maximum for randomly chosen n.batch.
unbalance.trt	Treatment for unbalanced labels.
bootstrap	Indicator for whether a bootstrap sampling is used.
prop.train	If bootstrap == FALSE, a training/validation cut for the data will be used (0, 1).
prop.keep	The proportion of DNNs to be kept in the ensemble.
best.opti	Whether to run the algorithm to keep the optimal subset of DNNs.
min.keep	Minimal number of DNNs to be kept.
verbose	Whether the progress to be printed.

Value

Returns a dnnetEnsemble object.

See Also

dnnetEnsemble-class
dnnetInput-class

SkadiEye/deepTL documentation built on Nov. 17, 2022, 1:41 p.m.