Learning Trees

• Trees:

  • At most one parent per variable

• Why trees?

  • Elegant math
    • we can solve the optimization problem
  • Sparse parameterization
    • avoid overfitting

Previous slide

Next slide

Back to first slide

View graphic version