Github user yanboliang commented on a diff in the pull request:
https://github.com/apache/spark/pull/12066#discussion_r61102661
--- Diff:
mllib/src/main/scala/org/apache/spark/ml/classification/Classifier.scala ---
@@ -115,7 +116,16 @@ abstract class ClassificationModel[FeaturesType, M <:
ClassificationModel[Featur
}
predictUDF(col(getFeaturesCol))
}
- outputData = outputData.withColumn(getPredictionCol, predUDF)
+ // determine number of classes either from metadata if provided.
+ val labelSchema = dataset.schema($(labelCol))
+ // extract label metadata from label column if present, or create a
nominal attribute
+ // to output the number of labels
+ val labelAttribute = Attribute.fromStructField(labelSchema) match {
--- End diff --
Reply:
1, I agree that it's better to extract labelAttribute during model fitting,
but it need to be transferred to model as an argument of model construction
which may take API breaking change. Is this reasonable? Actually
[```OneVsRestModel```](https://github.com/apache/spark/blob/master/mllib/src/main/scala/org/apache/spark/ml/classification/OneVsRest.scala#L136)
has already done similar work.
2, I referred that
[```OneVsRest```](https://github.com/apache/spark/blob/master/mllib/src/main/scala/org/apache/spark/ml/classification/OneVsRest.scala#L347)
and found the attribute was named as ```label```. May be we should rename it
as well.
3, Agree.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
