predict.swag: Predict method for SWAG

In swag: Sparse Wrapper Algorithm

Description

Gives predictions for different train learners obtained by swag.

Usage

## S3 method for class 'swag'
predict(
  object,
  newdata = NULL,
  type = c("best", "cv_performance", "attribute"),
  cv_performance = NULL,
  attribute = NULL,
  ...
)

Arguments

object	An object of class swag.
newdata	an optional set of data to predict on. If NULL the original training data are used.
type	type of prediction required. The default is "best", it takes the best model (with lowest CV errors). The option "cv_performance" (which requires cv_performance) allows to set a level of CV errors under which models are predicted. The option "attribute" (which requires attribute) allows to specify an attribute at which models are predicted.
cv_performance	a level of CV errors (between 0 and 1) combines with type "cv_performance".
attribute	an attribute combines with type "attribute".
...	Not used for the moment.

Details

Currently the different train learners are trained (again) to make the predictions.

Value

Predictions .

Author(s)

Gaetan Bakalli, Samuel Orso and Cesare Miglioli

swag documentation built on Nov. 10, 2020, 3:52 p.m.