SLScreenExtra: SLScreenExtra: A collection of screening algorithms for...

In saraemoore/SLScreenExtra: A Collection of Additional Feature Selection Algorithms and Utilities for SuperLearner

SLScreenExtra: A collection of screening algorithms for SuperLearner

Description

The SLScreenExtra package provides feature selection or screen.*() functions that plug into the R package for the ensemble learner SuperLearner. Some of the screening algorithms (screen.wgtd.*()) in this package utilize observation weights (obsWeights) in their calculations.

Filter methods

• Correlation:

  • screen.wgtd.corP

  • screen.wgtd.corRank

  • screen.FSelector.correlation

• Comparison test:

  • screen.wgtd.ttest

  • screen.wgtd.ttestP

  • screen.FSelector.chi.squared

  • screen.FSelector.oneR

• screen.FSelector.entropy

• screen.FSelector.relief

Embedded methods

• Regularized/penalized regression:

  • screen.wgtd.lasso

  • screen.wgtd.elasticnet

• Multivariate adaptive regression splines:

  • screen.earth

• Random decision forests:

  • screen.randomForest.imp

  • screen.ranger

  • screen.FSelector.random.forest.importance

Feature subset space search methods

• screen.FSelector.cfs

• screen.FSelector.consistency

Utility methods

• Produce tidy output from a SuperLearner or CV.SuperLearner object:

  • tidy.SuperLearner

  • tidy.CV.SuperLearner

saraemoore/SLScreenExtra documentation built on Nov. 4, 2023, 9:31 p.m.