Description
The
{CrossValidator}in Scala supports
{avgMetrics}since 1.5.0, which allows the user to evaluate how well each
{ParamMap}in the grid search performed and identify the best parameters. We should support this in PySpark as well.
Attachments
Issue Links
- is duplicated by
-
SPARK-11223 PySpark CrossValidatorModel does not output metrics for every param in paramGrid
- Resolved
- is related to
-
SPARK-11937 Python API coverage check found issues for ML during 1.6 QA
- Closed
- relates to
-
SPARK-14978 PySpark TrainValidationSplitModel should support validationMetrics
- Resolved
- links to