When creating a training dataset using the Databricks Feature Store, a data scientist wants to ensure that features reflect only the values available at the time each label observation was recorded. Which Feature Store capability addresses this?

Question

Accepted Answer

B. Point-in-time feature joins. Point-in-time correctness is a Feature Store capability that ensures training datasets reflect the feature values that were available at the exact timestamp of each label observation. This prevents data leakage from future feature values.

When creating a training dataset using the Databricks Feature Store, a data scientist wants to ensure that features reflect only the values available at the time each label observation was recorded. Which Feature Store capability addresses this?

Related Questions

Discussion