SPARK-5020 [MLlib] GaussianMixtureModel.predictMembership() should take an RDD only

[tgaloppo]

Removed unnecessary parameters to predictMembership()

CC: @jkbradley

[AmplabJenkins]

Can one of the admins verify this patch?

[tgaloppo]

My repo seems to have some lingering commit tags.

[jkbradley]

I should have been clearer; I meant for predictMembership to be public. I thought perhaps the implementation would become of a private method, but I realize now it is fine leaving the code within predictMembership alone. Once the private modifier is removed, I think this will be good to go.
Thanks!

[jkbradley]

About the commit tags, I'm not quite sure how to remove them. Does rebasing fix it? If not, then you could checkout master again and copy the few changes in (and then do a force push to your PR branch).

[tgaloppo]

@jkbradley No, private modifier was a gaffe on my part. This is corrected.

I think I have corrected the lingering commits... rebase did not work for me, but a hard reset seems to have done the trick.

[SparkQA]

Test build #558 has started for PR 3854 at commit 0f1d96e.

• This patch merges cleanly.

[jkbradley]

@tgaloppo Ok, thanks for the fix! LGTM once the tests are done.

@tgaloppo @mengxr As long as this method is being edited, do you like the name predictMembership for soft clustering? I assume we may eventually use the same term for other clustering methods, includes LDA.

[tgaloppo]

@jkbradley I am not crazy about the name predictMembership() ... to me it implies the hard assignment; a simple change like predictMemberships() might be more clear, or predictSoft(), or (thinking from a slightly different direction) allocate(). Any of those should be robust enough to reuse for soft k-means or LDA (or other such partial assignment algorithms).

[mengxr]

It may be hard for users to tell the difference between predict and predictMembership, because predict is also predicting the membership. predictFuzzy or predictSoft sounds better to me.

[jkbradley]

+1 for predictSoft

[SparkQA]

Test build #558 has finished for PR 3854 at commit 0f1d96e.

• This patch fails Spark unit tests.
• This patch merges cleanly.
• This patch adds no public classes.

[jkbradley]

streaming failure

[tgaloppo]

Ok, I will rename the method to predictSoft()

Not sure what to make of the streaming failure ??

[mengxr]

add to whitelist

[mengxr]

test this please

[SparkQA]

Test build #24971 has started for PR 3854 at commit 1bf4669.

• This patch merges cleanly.

[jkbradley]

@tgaloppo The streaming failure is unrelated; it's an unreliable test (which is being fixed). Nothing for you to do.

[SparkQA]

Test build #24971 has finished for PR 3854 at commit 1bf4669.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds no public classes.

[AmplabJenkins]

Test PASSed.
Refer to this link for build results (access rights to CI server needed):
https://amplab.cs.berkeley.edu/jenkins//job/SparkPullRequestBuilder/24971/
Test PASSed.

[mengxr]

LGTM. Merged into master. Thanks!