We used the name `Dataset` to refer to `SchemaRDD` in 1.2 in ML pipelines and created this example file. Since `Dataset` has a new meaning in Spark 1.6, we should rename it to avoid confusion. This PR also removes support for dense format to simplify the example code.

cc: yinxusen

Author: Xiangrui Meng <meng@databricks.com>

Closes #9873 from mengxr/SPARK-11895.