def test_grid_from_X():
# tests for _grid_from_X: sanity check for output, and for shapes.
# Make sure that the grid is a cartesian product of the input (it will use
# the unique values instead of the percentiles)
percentiles = (.05, .95)
grid_resolution = 100
X = np.asarray([[1, 2], [3, 4]])
grid, axes = _grid_from_X(X, percentiles, grid_resolution)
assert_array_equal(grid, [[1, 2], [1, 4], [3, 2], [3, 4]])
assert_array_equal(axes, X.T)
# test shapes of returned objects depending on the number of unique values
# for a feature.
rng = np.random.RandomState(0)
grid_resolution = 15
# n_unique_values > grid_resolution
X = rng.normal(size=(20, 2))
grid, axes = _grid_from_X(X, percentiles, grid_resolution=grid_resolution)
assert grid.shape == (grid_resolution * grid_resolution, X.shape[1])
assert np.asarray(axes).shape == (2, grid_resolution)
# n_unique_values < grid_resolution, will use actual values
n_unique_values = 12
X[n_unique_values - 1:, 0] = 12345
rng.shuffle(X) # just to make sure the order is irrelevant
grid, axes = _grid_from_X(X, percentiles, grid_resolution=grid_resolution)
assert grid.shape == (n_unique_values * grid_resolution, X.shape[1])
# axes is a list of arrays of different shapes
assert axes[0].shape == (n_unique_values, )
assert axes[1].shape == (grid_resolution, )
def _grid(self, X, features, percentiles, grid_resolution):
features_indices = np.asarray(
_get_column_indices(X, features), dtype=np.int32, order='C'
).ravel()
grid, values = _grid_from_X(
_safe_indexing(X, features_indices, axis=1), percentiles,
grid_resolution
)
return grid, values
def partial_dependence(estimator, X, features):
"""Calculate the partial dependence of features.
Partial dependence of a feature (or a set of features) corresponds
to the average response of an estimator for each possible value of
the feature. Code from sklearn's _partial_dependence.py. Note that
this implementation always uses method="brute", grid_resolution=100
and percentiles=(0.05, 0.95).
Parameters
----------
estimator : class
Model estimator to use.
X : pd.DataFrame
Feature set used to generate a grid of values for the target
features (where the partial dependence is evaluated), and
also to generate values for the complement features.
features : int, str or sequence
The feature or pair of interacting features for which the
partial dependency should be computed.
Returns
-------
averaged_predictions: np.ndarray
Average of the predictions.
values: list
Values used for the predictions.
"""
grid, values = _grid_from_X(_safe_indexing(X, features, axis=1),
(0.05, 0.95), 100)
averaged_predictions, _ = _partial_dependence_brute(
estimator, grid, features, X, "auto")
# Reshape averaged_predictions to (n_outputs, n_values_feature_0, ...)
averaged_predictions = averaged_predictions.reshape(
-1, *[val.shape[0] for val in values])
return averaged_predictions, values
def test_grid_from_X_error(grid_resolution, percentiles, err_msg):
X = np.asarray([[1, 2], [3, 4]])
with pytest.raises(ValueError, match=err_msg):
_grid_from_X(X,
grid_resolution=grid_resolution,
percentiles=percentiles)
