Erik Erlandson created SPARK-5876:
-------------------------------------
Summary: generalize the type of categoricalFeaturesInfo to
PartialFunction[Int, Int]
Key: SPARK-5876
URL: https://issues.apache.org/jira/browse/SPARK-5876
Project: Spark
Issue Type: Improvement
Components: MLlib
Reporter: Erik Erlandson
Priority: Minor
The decision tree training takes a parameter {{categoricalFeaturesInfo}} of
type {{Map\[Int,Int\]}} that encodes information about any features that are
categories and how many categorical values are present.
It would be useful to generalize this type to its superclass
{{PartialFunction\[Int,Int\]}}, which would be backward compatible with
{{Map\[Int,Int\]}}, but can also accept a {{Seq\[Int\]}}, or any other partial
function.
Would need to verify that any tests for key definition in the mapping are using
{{isDefinedAt(key)}} instead of {{contains(key)}}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
