ForwardSelection: Implement forward feature selection and return data with...

In UBC-MDS/rb4model: rb4model

Description

Implement forward feature selection and return data with selected features Uses root mean squared error for regression and accuracy for classification

Usage

ForwardSelection(
  my_mod,
  feature,
  label,
  max_f = NA,
  type = "classification",
  cv = 3
)

Arguments

my_mod	model name in string (must be in caret::modelLookup())
feature	training dataset with features
label	training dataset with labels.
max_f	maximum amount of features to select
type	problem type. (Must be 'regression' or 'classification')
cv	number of folds for cross validation

Value

The dataset with selected features.

Examples

y <- iris$Species
x <- iris[c(1,2,3,4)]
ffs <- ForwardSelection(feature=x, label=y, my_mod="rf")
print(x[ffs])

UBC-MDS/rb4model documentation built on March 31, 2020, 12:54 a.m.