Practical insights

Results

8 results (published only)

• Feature engineering and preprocessing

  Build target encodings in time order on time-stamped data, because random cross fitting borrows from the future

  A table of time-stamped outcomes, such as no-shows by departure date, is target encoded with a standard k-fold cross fit and then judged on a later period.

  hardpractitioner heuristic

• Feature engineering and preprocessing

  Shrink a thin route's target encoding toward its origin mean before the global mean

  Route-level target encoding is shrunk toward the overall rate, yet routes sharing an origin behave alike and new routes keep appearing with a handful of flights.

  hardpractitioner heuristic

• Feature engineering and preprocessing

  Bracket a suspected MNAR column with a delta adjustment instead of arguing about MAR versus MNAR

  A decision-critical column, such as stated income or a survey score, may be missing because of its own value, and the team debates the mechanism from missingness plots.

  harddocumented fact

• Feature engineering and preprocessing

  Replay logged arrival times to mask training features the way serving will see them

  A feature comes from a feed that lands hours after the event; the warehouse history is complete, but a live scoring request often lacks the value.

  hardpractitioner heuristic

• Feature engineering and preprocessing

  Include the outcome when imputing for an inferential model and leave it out of a deployed imputer

  You must impute predictors, and the same dataset serves both a coefficient-interpretation study and a model to be deployed for scoring.

  harddocumented fact

• Feature engineering and preprocessing

  Treat a feature's time-to-live as a modelling choice, since it decides which training rows receive nulls

  A feature store applies a time-to-live to a feature view, and training sets are built with point-in-time joins over entities with irregular update gaps.

  harddocumented fact

• Feature engineering and preprocessing

  Add a train-versus-live classifier to catch feature drift that per-feature tests miss

  Every per-feature drift test is quiet, yet model quality is slipping, and the features are strongly related to one another.

  hardpractitioner heuristic

• Feature engineering and preprocessing

  Keep only features that survive repeated subsampling when selection is noisy

  A lasso or importance-based selector returns a different shortlist when a few hundred rows change, and the team has to freeze a feature set.

  harddocumented fact