In an ecommerce use case, let's say product clicked=1, add to cart=2, purchased=3 are my labels. Does it make sense to also use those as features? Or does it not matter? I also have many other features such as bm25, price, fast delivery, star rating, etc.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
In an ecommerce use case, let's say product clicked=1, add to cart=2, purchased=3 are my labels. Does it make sense to also use those as features? Or does it not matter? I also have many other features such as bm25, price, fast delivery, star rating, etc.
I can't say I've seen the labels as being used as features before, although how you are defining them seem to be useful as features.
But using a label value as a feature to predict label value seems a bit problematic.
The other feature candidates you mention seem perfectly valid as predictors, and I'd definitely make use of them.
I don't think I'd use the labels as features, but you can certainly try them to see how the resulting models fare against validation and test data.
SMH