Skip to content

[MRG+1] split data using _safe_split in _permutaion_test_scorer to fix error… #5697

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.

Sign up for GitHub

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

Merged
merged 1 commit into from
Dec 29, 2016
Merged

[MRG+1] split data using _safe_split in _permutaion_test_scorer to fix error… #5697

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 5 additions & 3 deletions sklearn/cross_validation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1756,8 +1756,10 @@ def _permutation_test_score(estimator, X, y, cv, scorer):
"""Auxiliary function for permutation_test_score"""
avg_score = []
for train, test in cv:
estimator.fit(X[train], y[train])
avg_score.append(scorer(estimator, X[test], y[test]))
X_train, y_train = _safe_split(estimator, X, y, train)
X_test, y_test = _safe_split(estimator, X, y, test, train)
estimator.fit(X_train, y_train)
avg_score.append(scorer(estimator, X_test, y_test))
return np.mean(avg_score)


Expand All @@ -1770,7 +1772,7 @@ def _shuffle(y, labels, random_state):
for label in np.unique(labels):
this_mask = (labels == label)
ind[this_mask] = random_state.permutation(ind[this_mask])
return y[ind]
return safe_indexing(y, ind)


def check_cv(cv, X=None, y=None, classifier=False):
Expand Down
8 changes: 5 additions & 3 deletions sklearn/model_selection/_validation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -622,8 +622,10 @@ def _permutation_test_score(estimator, X, y, groups, cv, scorer):
"""Auxiliary function for permutation_test_score"""
avg_score = []
for train, test in cv.split(X, y, groups):
estimator.fit(X[train], y[train])
avg_score.append(scorer(estimator, X[test], y[test]))
X_train, y_train = _safe_split(estimator, X, y, train)
X_test, y_test = _safe_split(estimator, X, y, test, train)
estimator.fit(X_train, y_train)
avg_score.append(scorer(estimator, X_test, y_test))
return np.mean(avg_score)


Expand All @@ -636,7 +638,7 @@ def _shuffle(y, groups, random_state):
for group in np.unique(groups):
this_mask = (groups == group)
indices[this_mask] = random_state.permutation(indices[this_mask])
return y[indices]
return safe_indexing(y, indices)


def learning_curve(estimator, X, y, groups=None,
Expand Down
19 changes: 19 additions & 0 deletions sklearn/model_selection/tests/test_validation.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -966,3 +966,22 @@ def test_score_memmap():
break
except WindowsError:
sleep(1.)


def test_permutation_test_score_pandas():
# check permutation_test_score doesn't destroy pandas dataframe
types = [(MockDataFrame, MockDataFrame)]
try:
from pandas import Series, DataFrame
types.append((Series, DataFrame))
except ImportError:
pass
for TargetType, InputFeatureType in types:
# X dataframe, y series
iris = load_iris()
X, y = iris.data, iris.target
X_df, y_ser = InputFeatureType(X), TargetType(y)
check_df = lambda x: isinstance(x, InputFeatureType)
check_series = lambda x: isinstance(x, TargetType)
clf = CheckingClassifier(check_X=check_df, check_y=check_series)
permutation_test_score(clf, X_df, y_ser)
Copy link
Member

Choose a reason for hiding this comment

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

Ideally we should check that, for fixed random_state, results are identical regardless of input type.

Copy link
Contributor Author

Choose a reason for hiding this comment

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

At the moment, such input tests are also not present for the other validation methods, there is only: test_cross_val_score_pandas(), test_cross_val_predict_pandas().

I wonder is that really a necessity?