[SPARK-11975] [ML] Remove duplicate mllib example (DT/RF/GBT in Java/Python)

[yanboliang]

Remove duplicate mllib example (DT/RF/GBT in Java/Python).
Since we have tutorial code for DT/RF/GBT classification/regression in Scala/Java/Python and example applications for DT/RF/GBT in Scala, so we mark these as duplicated and remove them.
@mengxr

[SparkQA]

Test build #46658 has finished for PR 9954 at commit 96a4527.

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

[SparkQA]

Test build #46659 has finished for PR 9954 at commit 416da6a.

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

[jkbradley]

Just one comment: I have gotten feedback from some people about using these scripts, so we might get some complaints after removing them. That said, I'm OK with removing them since command-line tools are not the best way to use Spark.

[yanboliang]

@jkbradley Users can use decision_tree_classification_example.py and decision_tree_regression_example.py instead to illustrate the usage of decision tree algorithm and model. I still keep the Scala example application (with command-line tools one) which has more complete input params. There is little difference between these two kinds of examples for Java/Python, so I think removing Java/Python ones is make sense.

[mengxr]

LGTM. Merged into master and branch-1.6. Thanks!