Fix QR based fitting #559
Merged
Fix QR based fitting #559
Conversation
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
andreasnoack
force-pushed
the
an/norank
branch
2 times, most recently
from
33d1584 to
35c3e94
Compare
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## master #559 +/- ##
==========================================
- Coverage 90.11% 89.95% -0.17%
==========================================
Files 8 8
Lines 1123 1105 -18
==========================================
- Hits 1012 994 -18
Misses 111 111 ☔ View full report in Codecov by Sentry. 🚨 Try these New Features:
|
andreasnoack
requested review from
devmotion
and removed request for
mousum-github
devmotion
reviewed
Sep 28, 2024
src/linpred.jl
Outdated
| end | ||
| p.delbeta[1:rnk] = R \ (p.qr.Q'*r̃)[1:rnk] |
There was a problem hiding this comment.
Maybe
Suggested change
| p.delbeta[1:rnk] = R \ (p.qr.Q'*r̃)[1:rnk] | |
| mul!(view(p.delbeta, 1:rnk), view(p.qr.Q, :, 1:rnk)', r̃) | |
| ldiv!(R, view(p.delbeta, 1:rnk)) |
?
There was a problem hiding this comment.
We can't make a view into Q, see above, but we can make a view into the result.
andreasnoack
commented
Nov 5, 2024
src/linpred.jl
Outdated
| end | ||
| p.delbeta[1:rnk] = R \ (p.qr.Q'*r̃)[1:rnk] |
There was a problem hiding this comment.
We can't make a view into Q, see above, but we can make a view into the result.
andreasnoack
force-pushed
the
an/norank
branch
4 times, most recently
from
37249f5 to
2b61e90
Compare
|
@devmotion I finally found a way to download the test dataset reliably so this one should be good to go. |
devmotion
reviewed
Nov 20, 2024
devmotion
approved these changes
Nov 20, 2024
Avoid erroring out for low rank design matrices when dropcollinear=false. Avoid unnecessary triangular solves. Avoid indexing in the Q. Avoid slicing R matrix in a way that triggers a minimum norm solution. Remove unnecessary temporaries.
andreasnoack
force-pushed
the
an/norank
branch
from
054f910 to
94f7198
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Avoid erroring out for low rank design matrices when dropcollinear=false.
Avoid unnecessary triangular solves. Avoid indexing in the Q.
Avoid slicing R matrix in a way that triggers a minimum norm solution.
Remove unnecessary temporaries.
Fixes #558
I removed the white space so please review with https://github.com/JuliaStats/GLM.jl/pull/559/files?w=1