spark-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Seth Hendrickson (JIRA)" <>
Subject [jira] [Commented] (SPARK-11861) Feature importances for decision trees
Date Fri, 20 Nov 2015 16:51:11 GMT


Seth Hendrickson commented on SPARK-11861:

It's not really different, except that when feature importance for random forests got implemented
it wasn't added to the decision tree API. This would simply add an API entry point in {{DecisionTreeRegressionModel}}
and {{DecisionTreeClassificationModel}}. Perhaps I'm missing your meaning?

Having a feature importance for Decision Trees will be useful when implementing a feature
importance for Gradient Boosted Trees according to  [SPARK-11730|],
since GBT feature importance would likely be some aggregation of feature importance for individual
trees in the ensemble. 

> Feature importances for decision trees
> --------------------------------------
>                 Key: SPARK-11861
>                 URL:
>             Project: Spark
>          Issue Type: New Feature
>          Components: ML
>            Reporter: Seth Hendrickson
>            Priority: Minor
> Feature importances should be added to decision trees leveraging the feature importance
implementation for Random Forests.

This message was sent by Atlassian JIRA

To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message