[SPARK-19939] [ML] Add support for association rules in ML

[huaxingao]

What changes were proposed in this pull request?

Adding support to Association Rules in Spark ml.fpm.

Why are the changes needed?

Support is an indication of how frequently the itemset of an association rule appears in the database and suggests if the rules are generally applicable to the dateset. Refer to wiki for more details.

Does this PR introduce any user-facing change?

Yes. Associate Rules now have support measure

How was this patch tested?

existing and new unit test

[SparkQA]

Test build #124392 has finished for PR 28903 at commit 1cc6560.

• This patch fails due to an unknown error code, -9.
• This patch merges cleanly.
• This patch adds no public classes.

[huaxingao]

retest this please

[SparkQA]

Test build #124399 has finished for PR 28903 at commit 1cc6560.

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

[huaxingao]

cc @srowen @zhengruifeng

[srowen]

I was going to say, can't a caller just compute this if desired? but I guess freqUnion isn't exposed?
If it's a simple addition to the returned DataFrame to expose it and it isn't otherwise available, I could see it.

[zhengruifeng]

Does it means that the support is computed in the .mllib side but not exposed in the .ml side?

[huaxingao]

freqUnion is computed in mllib side but not exposed.

[SparkQA]

Test build #124452 has finished for PR 28903 at commit c80ec10.

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

[srowen]

Merged to master

[huaxingao]

Thanks! @srowen @zhengruifeng