What are some pros and cons of Discriminant Analysis?

Pros:

  • Well-designed for multiclass problems
  • Closed form solution, meaning computationally efficient
  • Outputs class probabilities for each label
  • Does not require much hyperparameter tuning

Cons:

  • Assumes multivariate normality among features, though it is relatively robust to this assumption
  • Not as well suited for mixed data types as decision tree based methods