Uploaded image for project: 'Spark'
  1. Spark
  2. SPARK-24101

MulticlassClassificationEvaluator should use sample weight data

    XMLWordPrintableJSON

Details

    • Improvement
    • Status: Resolved
    • Major
    • Resolution: Fixed
    • 2.0.2
    • 3.0.0
    • ML
    • None

    Description

      The LogisticRegression and LinearRegression models support training with a weight column, but the corresponding evaluators do not support computing metrics using those weights. This breaks model selection using CrossValidator.

      Attachments

        Issue Links

          Activity

            People

              imatiach Ilya Matiach
              imatiach Ilya Matiach
              Votes:
              0 Vote for this issue
              Watchers:
              3 Start watching this issue

              Dates

                Created:
                Updated:
                Resolved: