Github user MLnick commented on a diff in the pull request:
https://github.com/apache/spark/pull/11119#discussion_r56117376
--- Diff: mllib/src/main/scala/org/apache/spark/ml/clustering/KMeans.scala
---
@@ -258,6 +290,27 @@ class KMeans @Since("1.5.0") (
@Since("1.5.0")
def setSeed(value: Long): this.type = set(seed, value)
+ /** @group setParam */
+ @Since("2.0.0")
+ def setInitialModel(value: KMeansModel): this.type = set(initialModel,
value)
--- End diff --
I think that there are certain params that have to be the same (k for
example) where overriding doesn't even make sense. So we can either print a
warning and continue with initial model params, or throw an error.
However overriding convergence tolerance or iterations, say, should be
possible and is certainly a feasible and valid user situation. In this case
we should support it.
We can document the behaviour accordingly and it will be a bit different
for each model...
On Tue, 15 Mar 2016 at 07:06, jkbradley <[email protected]> wrote:
> In mllib/src/main/scala/org/apache/spark/ml/clustering/KMeans.scala
> <https://github.com/apache/spark/pull/11119#discussion_r56115607>:
>
> > @@ -258,6 +290,27 @@ class KMeans @Since("1.5.0") (
> > @Since("1.5.0")
> > def setSeed(value: Long): this.type = set(seed, value)
> >
> > + /** @group setParam */
> > + @Since("2.0.0")
> > + def setInitialModel(value: KMeansModel): this.type =
set(initialModel, value)
>
> I think that's the user's problem if they overwrite k or other parameters
> after setting the initial model. However, I'm Ok with addressing this in
> the follow-up PR for overriding the Params when initialModel is set.
>
> â
> You are receiving this because you were mentioned. Reply to this email
> directly or view it on GitHub: <a href="
> https://github.com/apache/spark/pull/11119/files#r56115607";>
> https://github.com/apache/spark/pull/11119/files#r56115607</a>
>
---
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 [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
