1. It includes all combination of features.
  2. Uses evaluation criteria like accuracy, precision and R-Squared
  3. This method iteratively adds or removes features which support the best performance of the model
  4. Consider the interaction between features. This makes this method to perform well and produce better results
  5. However, it is highly expensive. Fiter methods are cost effective

Methods available:

  • Forward Selection
  • Step wise Backward Elimination
  • Exhaustive Search