8000 Revert "[MRG] Implement fitting intercept with `sparse_cg` solver in … · xhluca/scikit-learn@d82ce4c · GitHub
[go: up one dir, main page]
Skip to content

Commit d82ce4c

Browse files
author
Xing
committed
Revert "[MRG] Implement fitting intercept with sparse_cg solver in Ridge regression (scikit-learn#13336)"
This reverts commit 3ed0547.
1 parent 6fa1a79 commit d82ce4c

File tree

3 files changed

+14
-79
lines changed

3 files changed

+14
-79
lines changed

‎doc/whats_new/v0.21.rst

Lines changed: 0 additions & 4 deletions
Original file line numberDiff line numberDiff line change
@@ -230,10 +230,6 @@ Support for Python 3.4 and below has been officially dropped.
230230
in version 0.21 and will be removed in version 0.23.
231231
:issue:`12821` by :user:`Nicolas Hug <NicolasHug>`.
232232

233-
- |Enhancement| `sparse_cg` solver in :class:`linear_model.ridge.Ridge`
234-
now supports fitting the intercept (i.e. ``fit_intercept=True``) when
235-
inputs are sparse . :issue:`13336` by :user:`Bartosz Telenczuk <btel>`
236-
237233
:mod:`sklearn.manifold`
238234
............................
239235

‎sklearn/linear_model/ridge.py

Lines changed: 8 additions & 65 deletions
Original file line numberDiff line numberDiff line change
@@ -33,31 +33,9 @@
3333
from ..exceptions import ConvergenceWarning
3434

3535

36-
def _solve_sparse_cg(X, y, alpha, max_iter=None, tol=1e-3, verbose=0,
37-
X_offset=None, X_scale=None):
38-
39-
def _get_rescaled_operator(X):
40-
41-
X_offset_scale = X_offset / X_scale
42-
43-
def matvec(b):
44-
return X.dot(b) - b.dot(X_offset_scale)
45-
46-
def rmatvec(b):
47-
return X.T.dot(b) - X_offset_scale * np.sum(b)
48-
49-
X1 = sparse.linalg.LinearOperator(shape=X.shape,
50-
matvec=matvec,
51-
rmatvec=rmatvec)
52-
return X1
53-
36+
def _solve_sparse_cg(X, y, alpha, max_iter=None, tol=1e-3, verbose=0):
5437
n_samples, n_features = X.shape
55-
56-
if X_offset is None or X_scale is None:
57-
X1 = sp_linalg.aslinearoperator(X)
58-
else:
59-
X1 = _get_rescaled_operator(X)
60-
38+
X1 = sp_linalg.aslinearoperator(X)
6139
coefs = np.empty((y.shape[1], n_features), dtype=X.dtype)
6240

6341
if n_features > n_samples:
@@ -348,25 +326,6 @@ def ridge_regression(X, y, alpha, sample_weight=None, solver='auto',
348326
-----
349327
This function won't compute the intercept.
350328
"""
351-
352-
return _ridge_regression(X, y, alpha,
353-
sample_weight=sample_weight,
354-
solver=solver,
355-
max_iter=max_iter,
356-
tol=tol,
357-
verbose=verbose,
358-
random_state=random_state,
359-
return_n_iter=return_n_iter,
360-
return_intercept=return_intercept,
361-
X_scale=None,
362-
X_offset=None)
363-
364-
365-
def _ridge_regression(X, y, alpha, sample_weight=None, solver='auto',
366-
max_iter=None, tol=1e-3, verbose=0, random_state=None,
367-
return_n_iter=False, return_intercept=False,
368-
X_scale=None, X_offset=None):
369-
370329
if return_intercept and sparse.issparse(X) and solver != 'sag':
371330
if solver != 'auto':
372331
warnings.warn("In Ridge, only 'sag' solver can currently fit the "
@@ -436,12 +395,7 @@ def _ridge_regression(X, y, alpha, sample_weight=None, solver='auto',
436395

437396
n_iter = None
438397
if solver == 'sparse_cg':
439-
coef = _solve_sparse_cg(X, y, alpha,
440-
max_iter=max_iter,
441-
tol=tol,
442-
verbose=verbose,
443-
X_offset=X_offset,
444-
X_scale=X_scale)
398+
coef = _solve_sparse_cg(X, y, alpha, max_iter, tol, verbose)
445399

446400
elif solver == 'lsqr':
447401
coef, n_iter = _solve_lsqr(X, y, alpha, max_iter, tol)
@@ -538,35 +492,24 @@ def fit(self, X, y, sample_weight=None):
538492
np.atleast_1d(sample_weight).ndim > 1):
539493
raise ValueError("Sample weights must be 1D array or scalar")
540494

541-
# when X is sparse we only remove offset from y
542495
X, y, X_offset, y_offset, X_scale = self._preprocess_data(
543496
X, y, self.fit_intercept, self.normalize, self.copy_X,
544-
sample_weight=sample_weight, return_mean=True)
497+
sample_weight=sample_weight)
545498

546499
# temporary fix for fitting the intercept with sparse data using 'sag'
547-
if (sparse.issparse(X) and self.fit_intercept and
548-
self.solver != 'sparse_cg'):
549-
self.coef_, self.n_iter_, self.intercept_ = _ridge_regression(
500+
if sparse.issparse(X) and self.fit_intercept:
501+
self.coef_, self.n_iter_, self.intercept_ = ridge_regression(
550502
X, y, alpha=self.alpha, sample_weight=sample_weight,
551503
max_iter=self.max_iter, tol=self.tol, solver=self.solver,
552504
random_state=self.random_state, return_n_iter=True,
553505
return_intercept=True)
554-
# add the offset which was subtracted by _preprocess_data
555506
self.intercept_ += y_offset
556507
else:
557-
if sparse.issparse(X):
558-
# required to fit intercept with sparse_cg solver
559-
params = {'X_offset': X_offset, 'X_scale': X_scale}
560-
else:
561-
# for dense matrices or when intercept is set to 0
562-
params = {}
563-
564-
self.coef_, self.n_iter_ = _ridge_regression(
508+
self.coef_, self.n_iter_ = ridge_regression(
565509
X, y, alpha=self.alpha, sample_weight=sample_weight,
566510
max_iter=self.max_iter, tol=self.tol, solver=self.solver,
567511
random_state=self.random_state, return_n_iter=True,
568-
return_intercept=False, **params)
569-
512+
return_intercept=False)
570513
self._set_intercept(X_offset, y_offset, X_scale)
571514

572515
return self

‎sklearn/linear_model/tests/test_ridge.py

Lines changed: 6 additions & 10 deletions
Original file line numberDiff line numberDiff line change
@@ -815,25 +815,21 @@ def test_n_iter():
815815
def test_ridge_fit_intercept_sparse():
816816
X, y = make_regression(n_samples=1000, n_features=2, n_informative=2,
817817
bias=10., random_state=42)
818-
819818
X_csr = sp.csr_matrix(X)
820819

821-
for solver in ['sag', 'sparse_cg']:
820+
for solver in ['saga', 'sag']:
822821
dense = Ridge(alpha=1., tol=1.e-15, solver=solver, fit_intercept=True)
823822
sparse = Ridge(alpha=1., tol=1.e-15, solver=solver, fit_intercept=True)
824823
dense.fit(X, y)
825-
with pytest.warns(None) as record:
826-
sparse.fit(X_csr, y)
827-
assert len(record) == 0
824+
sparse.fit(X_csr, y)
828825
assert_almost_equal(dense.intercept_, sparse.intercept_)
829826
assert_array_almost_equal(dense.coef_, sparse.coef_)
830827

831828
# test the solver switch and the corresponding warning
832-
for solver in ['saga', 'lsqr']:
833-
sparse = Ridge(alpha=1., tol=1.e-15, solver=solver, fit_intercept=True)
834-
assert_warns(UserWarning, sparse.fit, X_csr, y)
835-
assert_almost_equal(dense.intercept_, sparse.intercept_)
836-
assert_array_almost_equal(dense.coef_, sparse.coef_)
829+
sparse = Ridge(alpha=1., tol=1.e-15, solver='lsqr', fit_intercept=True)
830+
assert_warns(UserWarning, sparse.fit, X_csr, y)
831+
assert_almost_equal(dense.intercept_, sparse.intercept_)
832+
assert_array_almost_equal(dense.coef_, sparse.coef_)
837833

838834

839835
def test_errors_and_values_helper():

0 commit comments

Comments
 (0)
0