StackedEnsemble optimization: skip predictions for base models with coef 0 when using GLM metalearner

Description

During validation (StackedEnsemble) or scoring (StackedEnsembleModel), before generating predictions for a base model:

  1. check if metalearner thinks it's worth it:

    1. in case of GLM, if associated beta coef is not 0.

    2. for others return true.

  2. if not, skip this base model.

  3. apply same logic when generating SE mojo to exclude unnecessary base models.

Assignee

Tomas Fryda

Fix versions

None

Reporter

Sebastien Poirier

Support ticket URL

None

Labels

Affected Spark version

None

Customer Request Type

None

Task progress

None

CustomerVisible

No

Components

Priority

Major
Configure