Skip to content

Navigation Menu

Sign in
Appearance settings

Search code, repositories, users, issues, pull requests...

Provide feedback

We read every piece of feedback, and take your input very seriously.

Saved searches

Use saved searches to filter your results more quickly
Appearance settings

API Use grid_values instead of pdp_values in partial_dependence #25732

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 2 commits into from
Mar 2, 2023
Merged

API Use grid_values instead of pdp_values in partial_dependence #25732

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
6b10168
CLN Use grid_values instaed of pdp_values in partial_dependence
thomasjpfan Feb 28, 2023
3f23f24
DOC Adds PR number
thomasjpfan Feb 28, 2023
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
4 changes: 2 additions & 2 deletions 4 doc/whats_new/v1.3.rst
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -172,9 +172,9 @@ Changelog
.........................

- |API| :func:`inspection.partial_dependence` returns a :class:`utils.Bunch` with
new key: `pdp_values`. The `values` key is deprecated in favor of `pdp_values`
new key: `grid_values`. The `values` key is deprecated in favor of `grid_values`
and the `values` key will be removed in 1.5.
:pr:`21809` by `Thomas Fan`_.
:pr:`21809` and :pr:`25732` by `Thomas Fan`_.

:mod:`sklearn.linear_model`
...........................
Expand Down
16 changes: 8 additions & 8 deletions 16 sklearn/inspection/_partial_dependence.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -369,22 +369,22 @@ def partial_dependence(

.. deprecated:: 1.3
The key `values` has been deprecated in 1.3 and will be removed
in 1.5 in favor of `pdp_values`. See `pdp_values` for details
in 1.5 in favor of `grid_values`. See `grid_values` for details
about the `values` attribute.

pdp_values : seq of 1d ndarrays
grid_values : seq of 1d ndarrays
The values with which the grid has been created. The generated
grid is a cartesian product of the arrays in ``pdp_values``.
``len(pdp_values) == len(features)``. The size of each array
``pdp_values[j]`` is either ``grid_resolution``, or the number of
grid is a cartesian product of the arrays in ``grid_values`` where
``len(grid_values) == len(features)``. The size of each array
``grid_values[j]`` is either ``grid_resolution``, or the number of
unique values in ``X[:, j]``, whichever is smaller.

.. versionadded:: 1.3

``n_outputs`` corresponds to the number of classes in a multi-class
setting, or to the number of tasks for multi-output regression.
For classical regression and binary classification ``n_outputs==1``.
``n_values_feature_j`` corresponds to the size ``pdp_values[j]``.
``n_values_feature_j`` corresponds to the size ``grid_values[j]``.

See Also
--------
Expand Down Expand Up @@ -561,10 +561,10 @@ def partial_dependence(

msg = (
"Key: 'values', is deprecated in 1.3 and will be removed in 1.5. "
"Please use 'pdp_values' instead."
"Please use 'grid_values' instead."
)
pdp_results._set_deprecated(
values, new_key="pdp_values", deprecated_key="values", warning_message=msg
values, new_key="grid_values", deprecated_key="values", warning_message=msg
)

if kind == "average":
Expand Down
6 changes: 3 additions & 3 deletions 6 sklearn/inspection/_plot/partial_dependence.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1256,7 +1256,7 @@ def plot(
else:
pd_results_ = []
for kind_plot, pd_result in zip(kind, self.pd_results):
current_results = {"pdp_values": pd_result["pdp_values"]}
current_results = {"grid_values": pd_result["grid_values"]}

if kind_plot in ("individual", "both"):
preds = pd_result.individual
Expand All @@ -1274,7 +1274,7 @@ def plot(
# get global min and max average predictions of PD grouped by plot type
pdp_lim = {}
for kind_plot, pdp in zip(kind, pd_results_):
values = pdp["pdp_values"]
values = pdp["grid_values"]
preds = pdp.average if kind_plot == "average" else pdp.individual
min_pd = preds[self.target_idx].min()
max_pd = preds[self.target_idx].max()
Expand Down Expand Up @@ -1402,7 +1402,7 @@ def plot(
):
avg_preds = None
preds = None
feature_values = pd_result["pdp_values"]
feature_values = pd_result["grid_values"]
if kind_plot == "individual":
preds = pd_result.individual
elif kind_plot == "average":
Expand Down
8 changes: 4 additions & 4 deletions 8 sklearn/inspection/_plot/tests/test_plot_partial_dependence.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -103,7 +103,7 @@ def test_plot_partial_dependence(grid_resolution, pyplot, clf_diabetes, diabetes
target_idx = disp.target_idx

line_data = line.get_data()
assert_allclose(line_data[0], avg_preds["pdp_values"][0])
assert_allclose(line_data[0], avg_preds["grid_values"][0])
assert_allclose(line_data[1], avg_preds.average[target_idx].ravel())

# two feature position
Expand Down Expand Up @@ -243,7 +243,7 @@ def test_plot_partial_dependence_str_features(
assert line.get_alpha() == 0.8

line_data = line.get_data()
assert_allclose(line_data[0], avg_preds["pdp_values"][0])
assert_allclose(line_data[0], avg_preds["grid_values"][0])
assert_allclose(line_data[1], avg_preds.average[target_idx].ravel())

# contour
Expand Down Expand Up @@ -279,7 +279,7 @@ def test_plot_partial_dependence_custom_axes(pyplot, clf_diabetes, diabetes):
target_idx = disp.target_idx

line_data = line.get_data()
assert_allclose(line_data[0], avg_preds["pdp_values"][0])
assert_allclose(line_data[0], avg_preds["grid_values"][0])
assert_allclose(line_data[1], avg_preds.average[target_idx].ravel())

# contour
Expand Down Expand Up @@ -466,7 +466,7 @@ def test_plot_partial_dependence_multiclass(pyplot):
disp_target_0.pd_results, disp_symbol.pd_results
):
assert_allclose(int_result.average, symbol_result.average)
assert_allclose(int_result["pdp_values"], symbol_result["pdp_values"])
assert_allclose(int_result["grid_values"], symbol_result["grid_values"])

# check that the pd plots are different for another target
disp_target_1 = PartialDependenceDisplay.from_estimator(
Expand Down
28 changes: 14 additions & 14 deletions 28 sklearn/inspection/tests/test_partial_dependence.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -109,7 +109,7 @@ def test_output_shape(Estimator, method, data, grid_resolution, features, kind):
kind=kind,
grid_resolution=grid_resolution,
)
pdp, axes = result, result["pdp_values"]
pdp, axes = result, result["grid_values"]

expected_pdp_shape = (n_targets, *[grid_resolution for _ in range(len(features))])
expected_ice_shape = (
Expand Down Expand Up @@ -435,7 +435,7 @@ def test_partial_dependence_easy_target(est, power):
est, features=[target_variable], X=X, grid_resolution=1000, kind="average"
)

new_X = pdp["pdp_values"][0].reshape(-1, 1)
new_X = pdp["grid_values"][0].reshape(-1, 1)
new_y = pdp["average"][0]
# add polynomial features if needed
new_X = PolynomialFeatures(degree=power).fit_transform(new_X)
Expand Down Expand Up @@ -655,7 +655,7 @@ def test_partial_dependence_sample_weight():

pdp = partial_dependence(clf, X, features=[1], kind="average")

assert np.corrcoef(pdp["average"], pdp["pdp_values"])[0, 1] > 0.99
assert np.corrcoef(pdp["average"], pdp["grid_values"])[0, 1] > 0.99


def test_hist_gbdt_sw_not_supported():
Expand Down Expand Up @@ -693,8 +693,8 @@ def test_partial_dependence_pipeline():
)
assert_allclose(pdp_pipe["average"], pdp_clf["average"])
assert_allclose(
pdp_pipe["pdp_values"][0],
pdp_clf["pdp_values"][0] * scaler.scale_[features] + scaler.mean_[features],
pdp_pipe["grid_values"][0],
pdp_clf["grid_values"][0] * scaler.scale_[features] + scaler.mean_[features],
)


Expand Down Expand Up @@ -762,11 +762,11 @@ def test_partial_dependence_dataframe(estimator, preprocessor, features):
if preprocessor is not None:
scaler = preprocessor.named_transformers_["standardscaler"]
assert_allclose(
pdp_pipe["pdp_values"][1],
pdp_clf["pdp_values"][1] * scaler.scale_[1] + scaler.mean_[1],
pdp_pipe["grid_values"][1],
pdp_clf["grid_values"][1] * scaler.scale_[1] + scaler.mean_[1],
)
else:
assert_allclose(pdp_pipe["pdp_values"][1], pdp_clf["pdp_values"][1])
assert_allclose(pdp_pipe["grid_values"][1], pdp_clf["grid_values"][1])


@pytest.mark.parametrize(
Expand Down Expand Up @@ -797,7 +797,7 @@ def test_partial_dependence_feature_type(features, expected_pd_shape):
pipe, df, features=features, grid_resolution=10, kind="average"
)
assert pdp_pipe["average"].shape == expected_pd_shape
assert len(pdp_pipe["pdp_values"]) == len(pdp_pipe["average"].shape) - 1
assert len(pdp_pipe["grid_values"]) == len(pdp_pipe["average"].shape) - 1


@pytest.mark.parametrize(
Expand Down Expand Up @@ -851,17 +851,17 @@ def test_partial_dependence_bunch_values_deprecated():

msg = (
"Key: 'values', is deprecated in 1.3 and will be "
"removed in 1.5. Please use 'pdp_values' instead"
"removed in 1.5. Please use 'grid_values' instead"
)

with warnings.catch_warnings():
# Does not raise warnings with "pdp_values"
# Does not raise warnings with "grid_values"
warnings.simplefilter("error", FutureWarning)
pdp_values = pdp_avg["pdp_values"]
grid_values = pdp_avg["grid_values"]

with pytest.warns(FutureWarning, match=msg):
# Warns for "values"
values = pdp_avg["values"]

# "values" and "pdp_values" are the same object
assert values is pdp_values
# "values" and "grid_values" are the same object
assert values is grid_values
8 changes: 4 additions & 4 deletions 8 sklearn/utils/tests/test_bunch.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -12,16 +12,16 @@ def test_bunch_attribute_deprecation():
values = np.asarray([1, 2, 3])
msg = (
"Key: 'values', is deprecated in 1.3 and will be "
"removed in 1.5. Please use 'pdp_values' instead"
"removed in 1.5. Please use 'grid_values' instead"
)
bunch._set_deprecated(
values, new_key="pdp_values", deprecated_key="values", warning_message=msg
values, new_key="grid_values", deprecated_key="values", warning_message=msg
)

with warnings.catch_warnings():
# Does not warn for "pdp_values"
# Does not warn for "grid_values"
warnings.simplefilter("error")
v = bunch["pdp_values"]
v = bunch["grid_values"]

assert v is values

Expand Down
Morty Proxy This is a proxified and sanitized view of the page, visit original site.