Approach for Small Number of Items in Catalog

[chiragdaryani]

Hi, my training data involves implicit feedback from users, specifically their past purchase history (no ratings).

Our item catalog is small, around 20 items. The goal is to rank these items based on the probability of users making a purchase.

Considering the limited item size, I'm thinking of skipping the retrieval model and using a ranking model directly. Is this the right approach?

[rlcauvin]

We typically use the retrieval model to narrow the set of candidates to maybe 10 or 100 before ranking them with the ranking model, so yes, you can probably skip the retrieval stage.

[OmarMAmin]

Hi @chiragdaryani Are you providing the examples one by one? wondering if ranking can be improving things if the catalogue is only 20 items