Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

FIX Raise error when missing-values encountered in scikit-tree trees #264

Merged
merged 5 commits into from
May 6, 2024

Conversation

adam2392
Copy link
Collaborator

Fixes #263

Changes proposed in this pull request:

  • raises a ValueError when missing-value encountered in scikit-tree estimators

Before submitting

  • I've read and followed all steps in the Making a pull request
    section of the CONTRIBUTING docs.
  • I've updated or added any relevant docstrings following the syntax described in the
    Writing docstrings section of the CONTRIBUTING docs.
  • If this PR fixes a bug, I've added a test that will fail without my fix.
  • If this PR adds a new feature, I've added tests that sufficiently cover my new functionality.

After submitting

  • All GitHub Actions jobs for my pull request have passed.

Signed-off-by: Adam Li <adam2392@gmail.com>
Signed-off-by: Adam Li <adam2392@gmail.com>
Signed-off-by: Adam Li <adam2392@gmail.com>
@@ -64,3 +64,7 @@ def compute_similarity_matrix(self, X):
The similarity matrix among the samples.
"""
return compute_forest_similarity_matrix(self, X)

def _more_tags(self):
# XXX: no scikit-tree estimators support NaNs as of now

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Scikit-Tree appears to be based on Scikit-Learn 1.3.X, which supports missing values in standalone decision tree models (DecisionTreeClassifier, DecisionTreeRegressor) by default. The support for missing values in random forest models (RandomForestClassifier and RandomForestRegressor) was added in 1.4.X.

In that regard, perhaps the _more_tags method should be base class-dependent (one tag for DecisionTree and another one for RandomForest)?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Scikit-tree will depend on each version of scikit-learn successively very tightly for now, so we'll consider the next release of v0.8 to be tied to 1.4.x. Is that okay?

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Right now, when there is no missing value support, then it would be okay to attach the same "disable all"-type_more_tags method to the tree ABC, so that it affects all tree and tree ensemble estimators.

The exact SkTree <-> SkLearn version matching is not critical until SkTree really starts enabling missing values.

@adam2392
Copy link
Collaborator Author

I will merge this if LGTYou @vruusmann?

Signed-off-by: Adam Li <adam2392@gmail.com>
Signed-off-by: Adam Li <adam2392@gmail.com>
@adam2392
Copy link
Collaborator Author

CI failures seem unrelated: actions/setup-python#855

@adam2392 adam2392 merged commit dace5cf into main May 6, 2024
26 of 28 checks passed
@adam2392 adam2392 deleted the mv branch May 6, 2024 14:19
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

Current status of missing values (in oblique trees)
2 participants