dabl/dabl
Auf GitHub ansehenAdd target encoding for high cardinality categorical variables
Open
#222 geöffnet am 20. Feb. 2020
help wanted
Repository-Metriken
- Stars
- (728 Stars)
- PR-Merge-Metriken
- (Keine gemergten PRs in 30 T)
Beschreibung
We could take it either from dirty_cat: https://dirty-cat.github.io/stable/ or from category_encoders: https://contrib.scikit-learn.org/categorical-encoding/
I'm tending towards a cross-validation based approach, not loo, but smoothing might also work.