XGBoost: add support for "eta decay"

Description

Context:
AutoML is currently spending most of its time on exploration, and even though the learning_rate is sometimes too high or too low: cf. https://0xdata.atlassian.net/browse/PUBDEV-6293

For the exploitation part, instead of implementing our own iterator where we would train multiple time the same model, each time lowering the learning rate, which requires a lot of computation time, AutoML would benefit from XGBoost support for eta_decay (aka learn_rate_annealing in GBM).

Assignee

Michal Kurka

Fix versions

None

Reporter

Sebastien Poirier

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