Skip to content

mlpapers/feature-selection

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

53 Commits
 
 
 
 

Repository files navigation

Feature selection (variable selection)

Feature selection is the process of selecting a subset of relevant features (variables, predictors) for use in model construction (Wikipedia)

Why feature selection?

  1. Data exploration
  2. Curse of dimensionality
  3. Less features - faster models
  4. Better metrics

Filter methods

Filter methods use model-free ranking to filter less relevant features

Wrapper methods

Wrapper methods use a model and its performance to find the best feature subset

Embedded methods

Unsupervised and semi-supervised feature selection

Stable feature selection

Domain-specific

Meta feature selection

Packages

  • R
    • Package: fscaret (CRAN) Jakub Szlek
    • Package: praznik (Code) Miron Kursa
    • Package: FSinR (CRAN, Paper) Francisco Aragón-Royón, Alfonso Jiménez-Vílchez, Antonio Arauzo-Azofra, José Manuel Benítez
    • Package: VSURF (CRAN, Paper)
    • Package: spikeSlabGAM (Code, CRAN, Paper)
    • Package: copent (CRAN, Code, Paper)
  • Python
  • Julia
    • the main packages for ML in Julia are MLJ and Flow