pyensemble/wildwood

Put back `min_sample_split` parameter

Closed this issue · 1 comments

stephanegaiffas commented

This should :

  • Correctly set the option in the tree_context
  • This corresponds to the minimum number of both training and validation samples in a node
  • Use it in _grow for the tests and also in find_best_split_along_feature * in the tests
stephanegaiffas commented

Fixed by #57