Skip to content

Fix for ridge regression with sparse matrix input #2552

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

Closed
wants to merge 1 commit into from
Closed

Fix for ridge regression with sparse matrix input #2552

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
6 changes: 3 additions & 3 deletions sklearn/linear_model/ridge.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -645,8 +645,8 @@ def _values(self, alpha, y, v, Q, QT_y):
return y - (c / G_diag), c

def _pre_compute_svd(self, X, y):
if sparse.issparse(X):
raise TypeError("SVD not supported for sparse matrices")
if sparse.issparse(X) and hasattr(X, 'toarray'):
X = X.toarray()
Copy link
Member

Choose a reason for hiding this comment

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

I'm -1 on this PR because densifying a sparse matrix behind the scenes is really not a good idea. I'd rather improve the error message to tell users to call X = X.toarray() in their script.

U, s, _ = linalg.svd(X, full_matrices=0)
v = s ** 2
UT_y = np.dot(U.T, y)
Expand Down Expand Up @@ -701,7 +701,7 @@ def fit(self, X, y, sample_weight=1.0):
with_sw = len(np.shape(sample_weight))

if gcv_mode is None or gcv_mode == 'auto':
if sparse.issparse(X) or n_features > n_samples or with_sw:
if n_features > n_samples or with_sw:
gcv_mode = 'eigen'
Copy link
Contributor Author

Choose a reason for hiding this comment

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

Here, sparse matrices with n_samp >> n_feat (say typical of the type produced by text feature problems) are sent to eigen, where a dense n_samp x n_samp is produced (X dot X.T), when dense matrices of this shape would have been sent to SVD (and in the past sparse were also).

else:
gcv_mode = 'svd'
Expand Down