Algorithm · Ebay · Medium
Given a target product ID targetId and a collection of user browsing sessions sessions, where each session is a list of product IDs viewed in that session. Count, across only the sessions that contain targetId, how many sessions each other product appears together with targetId. Count each product at most once per session (session-level deduplication). Then return the k products with the highest co-occurrence counts as recommendations. Rules: Only consider sessions that…
Checking your access…