Description
Currently there is no Python implementation for ml.attribute, so we can not use "Attribute" in ML pipeline. Some transformers need this feature such as VectorSlicer can take a subarray of the original features by specifying column names which should contains in the column Attribute.
Attachments
Issue Links
- duplicates
-
SPARK-8516 ML attribute API in PySpark
-
- Resolved
-
- Is contained by
-
SPARK-9663 ML Python API coverage issues found during 1.5 QA
-
- Resolved
-