The Discriminant function can be modified without changing the classification outcome.

  • Scaling by a positive constant
  • Adding a constant

  • Applying a function

These transformations are often used to simplify the mathematical calculations involved in classification. Where is a monotonically increasing function

Use cases

  • Improving the confidence of the decision by having a very high
  • Using in the context of a multi-class problem to normalize each classification.

Q

How can you simplify a discriminant function without changing the classification result?