Github user sethah commented on a diff in the pull request:
https://github.com/apache/spark/pull/15593#discussion_r87504228
--- Diff:
mllib/src/main/scala/org/apache/spark/ml/classification/LogisticRegression.scala
---
@@ -489,13 +485,14 @@ class LogisticRegression @Since("1.2.0") (
val initialCoefWithInterceptArray =
initialCoefficientsWithIntercept.toArray
--- End diff --
Are there other linear models that use a matrix of coefficients? I agree
that thinking about indexing when flattening a matrix into an array is a pain,
but I don't really see how column major is _more_ difficult than row major. The
only place this would make much difference is that we wouldn't have to modify
the L2 reg update logic and the initial coefficients, but we still need to swap
between layouts in every iteration. I guess I don't see that this is obviously
simpler.
And for the review of this PR, I think we can be confident of the
correctness due to the robustness of the LogisticRegression test suite - which
has extensive correctness tests.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
