rfeRF: rfeRF

In enriquea/feseR: Feature Selection workflow in R

Description

Recursive feature elimination (RFE) method wrapped with a Random Forest (RF) algorithm for feature importance evaluation.

Usage

rfeRF(
  features,
  class,
  number.cv = 10,
  group.sizes = c(1:10, seq(15, 100, 5)),
  metric = "Accuracy",
  verbose = TRUE,
  tolerance = 0
)

Arguments

features	A numeric matrix as input.
class	Response variable as numeric vector. It will be coerced to factor.
number.cv	Number of cross-validation folds (10 default). Used during training phase.
group.sizes	A numeric vector of integers corresponding to the number of features that should be retained.
metric	Metric to evaluate performance ('Accuracy' (default), 'Kappa' or 'ROC').
verbose	Make the output verbose.
tolerance	Allow tolerance for evaluation metric (Default zero).

Value

A list the elements. See rfe for more details.

enriquea/feseR documentation built on March 30, 2021, 4:12 p.m.