[MRG+1] Changing default model for IterativeImputer to BayesianRidge#13038
Merged
glemaitre merged 8 commits intoscikit-learn:iterativeimputerfrom Jan 24, 2019
Merged
Conversation
Member
|
Please update the documentation. |
Contributor
Author
|
Whoops, forgot. Thanks. |
jnothman
reviewed
Jan 23, 2019
Member
jnothman
left a comment
There was a problem hiding this comment.
I assume this does not also affect documentation in the user guide.
We probably should have had a test that would have broken with this change, i.e. checking the class of each predictor. We should make sure there is such a test at least for custom predictors to ensure no regressions....
Contributor
Author
|
Sorry, I don't follow. The test would ensure what is going on with custom predictors? |
Member
|
I meant like #13039, but also for predictor=None |
8 tasks
Contributor
Author
|
I think you mean this? |
Member
|
Yeah that sort of thing. Historically we should have tested the effect of
sample_posterior is what I really mean... But now it is more
straightforward.
|
glemaitre
requested changes
Jan 24, 2019
Member
glemaitre
left a comment
There was a problem hiding this comment.
I would refactor the test with the previous one. It would require just a single statement more.
Otherwise LGTM
sklearn/tests/test_impute.py
Outdated
| assert len(set(hashes)) == len(hashes) | ||
|
|
||
|
|
||
| def test_iterative_imputer_bayesianridge_default(): |
Member
There was a problem hiding this comment.
Would it be better to just make this case in the previous test
Member
There was a problem hiding this comment.
I meant:
@pytest.mark.parametrize(
"predictor",
[None, DummyRegressor(), BayesianRidge(), ARDRegression(), RidgeCV()]
)
def test_iterative_imputer_predictors(predictor):
rng = np.random.RandomState(0)
n = 100
d = 10
X = sparse_random_matrix(n, d, density=0.10, random_state=rng).toarray()
imputer = IterativeImputer(missing_values=0,
n_iter=1,
predictor=predictor,
random_state=rng)
imputer.fit_transform(X)
# check that types are correct for predictors
hashes = []
for triplet in imputer.imputation_sequence_:
expected_type = type(predictor) if predictor is not None else type(BayesianRidge())
assert isinstance(triplet.predictor, expected_type)
hashes.append(id(triplet.predictor))
# check that each predictor is unique
assert len(set(hashes)) == len(hashes)
glemaitre
approved these changes
Jan 24, 2019
Member
|
Merging when CI turn green |
Member
|
Thanks for the change |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
As discussed in #13026 It turns out having
RidgeCVas the default is problematic for reproducibility. This default should be gentler.Note, this may not pass tests until a more stable example is merged via #13026.
Paging @jnothman