Recommender Systems challenge at Polimi A.Y. 2023-2024, this is formally part of the Recommender Systems course by prof. Paolo Cremonesi. Link to the challenge
The application domain was book recommendation. The datasets that were provided contained interactions of users with books, in particular, if the user attributed to the book a rating of at least 4. The main goal of the competition was to discover which items (books) a user will interact with. Thus, the challenge consisted in a pure collaborative filtering problem, since no ICM was provided. The evaluation metric was Mean Average Precision @ 10.
- Ranked 8th in the private leaderboard and 5th in the public one, among 60+ teams
- MAP@10 - private: 0.14107
- MAP@10 - public: 0.14231