zhengruifeng commented on a change in pull request #28710:
URL: https://github.com/apache/spark/pull/28710#discussion_r438625969
##########
File path: project/MimaExcludes.scala
##########
@@ -49,7 +49,34 @@ object MimaExcludes {
//[SPARK-31840] Add instance weight support in LogisticRegressionSummary
// weightCol in
org.apache.spark.ml.classification.LogisticRegressionSummary is present only in
current version
-
ProblemFilters.exclude[ReversedMissingMethodProblem]("org.apache.spark.ml.classification.LogisticRegressionSummary.weightCol")
+
ProblemFilters.exclude[ReversedMissingMethodProblem]("org.apache.spark.ml.classification.LogisticRegressionSummary.weightCol"),
+
+ //[SPARK-31893] Add a generic ClassificationSummary trait
+
ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("org.apache.spark.ml.classification.LogisticRegressionTrainingSummary.org$apache$spark$ml$classification$ClassificationSummary$_setter_$org$apache$spark$ml$classification$ClassificationSummary$$multiclassMetrics_="),
+
ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("org.apache.spark.ml.classification.LogisticRegressionTrainingSummary.org$apache$spark$ml$classification$ClassificationSummary$$multiclassMetrics"),
+
ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("org.apache.spark.ml.classification.LogisticRegressionTrainingSummary.weightCol"),
+
ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("org.apache.spark.ml.classification.BinaryLogisticRegressionTrainingSummary.org$apache$spark$ml$classification$BinaryClassificationSummary$_setter_$org$apache$spark$ml$classification$BinaryClassificationSummary$$sparkSession_="),
+
ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("org.apache.spark.ml.classification.BinaryLogisticRegressionTrainingSummary.org$apache$spark$ml$classification$BinaryClassificationSummary$_setter_$org$apache$spark$ml$classification$BinaryClassificationSummary$$binaryMetrics_="),
+
ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("org.apache.spark.ml.classification.BinaryLogisticRegressionTrainingSummary.org$apache$spark$ml$classification$BinaryClassificationSummary$$binaryMetrics"),
+
ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("org.apache.spark.ml.classification.BinaryLogisticRegressionTrainingSummary.scoreCol"),
+
ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("org.apache.spark.ml.classification.BinaryLogisticRegressionTrainingSummary.org$apache$spark$ml$classification$BinaryClassificationSummary$$sparkSession"),
+
ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("org.apache.spark.ml.classification.BinaryLogisticRegressionTrainingSummary.org$apache$spark$ml$classification$ClassificationSummary$_setter_$org$apache$spark$ml$classification$ClassificationSummary$$multiclassMetrics_="),
+
ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("org.apache.spark.ml.classification.BinaryLogisticRegressionTrainingSummary.org$apache$spark$ml$classification$ClassificationSummary$$multiclassMetrics"),
+
ProblemFilters.exclude[InheritedNewAbstractMethodProblem]("org.apache.spark.ml.classification.BinaryLogisticRegressionTrainingSummary.weightCol"),
+
ProblemFilters.exclude[IncompatibleResultTypeProblem]("org.apache.spark.ml.classification.LogisticRegressionSummary.asBinary"),
Review comment:
> in multinomial case, summary is still an instance of
LogisticRegressionTrainingSummary
Is it expected?
I think a reasonable behavior should be:
for any classification model, a ClassificationSummary should be attached;
if and only if numClasses=2, then this summary is also a
BinaryClassificationSummary
----------------------------------------------------------------
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
