[SPARK-7290][ML] Add StringVectorizer

[viirya]

JIRA: https://issues.apache.org/jira/browse/SPARK-7290

[SparkQA]

Test build #32490 has finished for PR 6078 at commit 032d32c.

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

[viirya]

retest this please.

[SparkQA]

Test build #32498 has finished for PR 6078 at commit 032d32c.

• This patch fails MiMa tests.
• This patch merges cleanly.
• This patch adds the following public classes (experimental):
  • class StringVectorizer extends Estimator[StringVectorizerModel] with StringVectorizerBase

[viirya]

retest this please.

[SparkQA]

Test build #32502 has finished for PR 6078 at commit 032d32c.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds the following public classes (experimental):
  • class StringVectorizer extends Estimator[StringVectorizerModel] with StringVectorizerBase

[mengxr]

@viirya This is essentially StringIndexer + OneHotEncoder. So any ideas about code reuse?

[viirya]

@mengxr Updated.
But by reusing StringIndexer and OneHotEncoder, we introduce an intermediate column for StringIndexer's output. Now I use _ + outputColumnName as its column name. It is very rare, but still possibly conflicting with dataset's columns.

[SparkQA]

Test build #32575 has finished for PR 6078 at commit 6eaee99.

• This patch passes all tests.
• This patch merges cleanly.
• This patch adds the following public classes (experimental):
  • class StringVectorizer extends Estimator[StringVectorizerModel] with StringVectorizerBase