The goal of recipeselectors is to provide extra supervised feature selection steps to be used with the tidymodels recipes package.

The package is under development.

Installation

devtools::install_github("stevenpawley/recipeselectors")

Feature Selection Methods

The following feature selection methods are implemented:

  • step_select_infgain provides Information Gain feature selection. This step requires the FSelectorRcpp package to be installed.

  • step_select_mrmr provides maximum Relevancy Minimum Redundancy feature selection. This step requires the praznik package to be installed.

  • step_select_roc provides ROC-based feature selection based on each predictors’ relationship with the response outcomeas measured using a Receiver Operating Characteristic curve. Thanks to Max Kuhn, along with many other useful suggestions.

  • step_select_xtab provides feature selection using statistical association (also thanks to Max Kuhn).

  • step_select_vip provides model-based selection using feature importance scores or coefficients. This method allows a parsnip model specification to be used to select a subset of features based on the models’ feature importances or coefficients. See below for details. Note, that this step will eventually be deprecated in favor of separate steps that contain the specific models that are most commonly used for feature selection such as step_select_forests, step_select_tree and step_select_linear.

  • step_select_boruta provides a Boruta feature selection step.

  • step_select_carscore provides a CAR score feature selection step for regression models. This step requires the care package to be installed.

  • step_select_forests, step_select_tree, and step_select_linear provide model-based methods of selecting a subset of features based on the model’s feature importance scores or coefficients. These steps, and potential step_select_rules, step_select_boost will replace the step_select_vip method.

Under Development

Methods that are planned to be added:

  • Relief-based methods (CORElearn package)

  • Ensemble feature selection (EFS package)

Notes on Wrapper Feature Selection Methods

The focus of recipeselectors is to provide extra recipes for filter-based feature selection. A single wrapper method is also included using the variable importance scores of selected algorithms for feature selection.

The step_select_vip is designed to work with the parsnip package and requires a base model specification that provides a method of ranking the importance of features, such as feature importance scores or coefficients, with one score per feature. The base model is specified in the step using the model parameter.

A limitation is that the model used in the step_select_vip cannot be tuned. This step will be replaced by a more appropriate structure that allows both variable selection and tuning for specific model types.

The parsnip package does not currently contain a method of pulling feature importance scores from models that support them. The recipeselectors package provides a generic function pull_importances for this purpose that accepts a fitted parsnip model, and returns a tibble with two columns ‘feature’ and ‘importance’:

model <- boost_tree(mode = "classification") %>%
  set_engine("xgboost")

model_fit <- model %>% 
  fit(Species ~., iris)

pull_importances(model_fit)

Most of the models and ‘engines’ that provide feature importances are implemented. In addition, h2o models are supported using the h2oparsnip package. Use methods(pull_importances) to list models that are currently implemented. If need to pull the feature importance scores from a model that is not currently supported in this package, then you can add a class to the pull_importances generic function which returns a two-column tibble:

pull_importances._ranger <- function(object, scaled = FALSE, ...) {
  scores <- ranger::importance(object$fit)

  # create a tibble with 'feature' and 'importance' columns
  scores <- tibble::tibble(
    feature = names(scores),
    importance = as.numeric(scores)
  )

  # optionally rescale the importance scores
  if (scaled)
    scores$importance <- scales::rescale(scores$importance)
  scores
}

An example of using the step_importance function:

library(parsnip)
library(recipes)
library(magrittr)

# load the example iris dataset
data(iris)

# define a base model to use for feature importances
base_model <- rand_forest(mode = "classification") %>%
  set_engine("ranger", importance = "permutation")

# create a preprocessing recipe
rec <- iris %>%
recipe(Species ~ .) %>%
step_select_vip(all_predictors(), model = base_model, top_p = 2,
                outcome = "Species")

prepped <- prep(rec)

# create a model specification
clf <- decision_tree(mode = "classification") %>%
set_engine("rpart")

clf_fitted <- clf %>%
  fit(Species ~ ., juice(prepped))