Settings: Drift Detection¶
H2O Model Validation offers an array of settings for a Drift Detection test. Below, each setting is described in turn.
Dataset¶
Defines the training dataset that H2O Model Validation uses as the base for the validation test to detect feature drifts.
Reference dataset¶
Defines the reference dataset that H2O Model Validation assesses for feature drifts. H2O Model Validation uses the training and reference dataset to detect feature drifts using the Population Stability Index (PSI).
Columns to drop¶
Defines the columns H2O Model Validation drops during the validation test. Typically drop columns refer to columns that can indicate a drift without an impact on the model, like columns not used by the model, record IDs, time columns, etc.
- Submit and view feedback for this page
- Send feedback about H2O Model Validation to cloud-feedback@h2o.ai