medium10 pts · first answer countsfeature-engineeringdata-leakage

Target encoding without leakage

You want to replace a high-cardinality category with the mean target for that category. How should you compute it for the training rows?

A.Use category means from the full dataset, test includedB.Use each row's own label as the feature valueC.Sort the categories alphabetically and number themD.Use out-of-fold category means with smoothing

Discussion 0

Join free to post in the discussion.

Target encoding without leakage

Discussion 0

Related questions