Reorder the algorithms trained in AutoML, prioritizing hardcoded XGBoost models

Description

On situations where we only are given enough time to train one model (currently a default RF), the results are sub-par. We are going to start instead with the group of hardcoded XGBoosts since they often do better than the default RF. This change should also be documented in the first bullet of the AutoML FAQ.

New order:

defaultXGBoosts();
defaultSearchGLM();
defaultRandomForest();
defaultGBMs();
defaultDeepLearning();
defaultExtremelyRandomTrees();
defaultSearchXGBoost();
defaultSearchGBM();
defaultSearchDL();
defaultStackedEnsembles();

Assignee

Sebastien Poirier

Fix versions

Reporter

Erin LeDell

Support ticket URL

None

Labels

None

Affected Spark version

None

Customer Request Type

None

Task progress

None

ReleaseNotesHidden

None

CustomerVisible

No

Components

Priority

Major
Configure