Uploaded image for project: 'Spark'
  1. Spark
  2. SPARK-26721

Bug in feature importance calculation in GBM (and possibly other decision tree classifiers)

    XMLWordPrintableJSON

    Details

    • Type: Bug
    • Status: Resolved
    • Priority: Major
    • Resolution: Fixed
    • Affects Version/s: 2.4.0
    • Fix Version/s: 3.0.0
    • Component/s: ML
    • Labels:
      None

      Description

      The feature importance calculation in org.apache.spark.ml.classification.GBTClassificationModel.featureImportances follows a flawed implementation from scikit-learn resulting in incorrect importance values. This error was recently discovered and updated in scikit-learn version 0.20.0. This error is inherited in the spark implementation and needs to be fixed here as well.

      As described in the scikit-learn release notes (https://scikit-learn.org/stable/whats_new.html#version-0-20-0):

      Fix Fixed a bug in ensemble.GradientBoostingRegressor and ensemble.GradientBoostingClassifier to have feature importances summed and then normalized, rather than normalizing on a per-tree basis. The previous behavior over-weighted the Gini importance of features that appear in later stages. This issue only affected feature importances. #11176 by Gil Forsyth.

      Full discussion of this error and debate ultimately validating the correctness of the change can be found in the comment thread of the scikit-learn pull request: https://github.com/scikit-learn/scikit-learn/pull/11176 

       

      I believe the main change required would be to the featureImportances function in mllib/src/main/scala/org/apache/spark/ml/tree/treeModels.scala , however, I do not have the experience to make this change myself.

        Attachments

          Issue Links

            Activity

              People

              • Assignee:
                mgaido Marco Gaido
                Reporter:
                danjump Daniel Jumper
              • Votes:
                0 Vote for this issue
                Watchers:
                4 Start watching this issue

                Dates

                • Created:
                  Updated:
                  Resolved: