Hacker News Reader

Normalization is a model-dependent transformation and happens after the feature store - needs to be consistent between training and inference pipelines.

View on HN

Normalization is overloaded. I was referring to schema normalization (3NF etc) not feature normalization - like standard scaling etc.

Ok, but star schema is denormalized. Snowflake is normalized.

To be pedantic, even in star schema - the dim tables are denormalized, fact tables are not.

I agree that my statement would be much better if used snowflake schema instead.

Reply on news.ycombinator.com