| Popis: |
A 3 month long RecSys 2020 challenge1 was organized by Twitter[1]. Various kinds of implicit feedback with varying levels of sparsity were considered. In subsequent sections, I describe my approach to address this task, which helped me land on the 4th place on the final leaderboard with the final score of 20. My team name was learner_recsys and link to my code is present on GitHub2. This is detailed in Table 1. I used a combination of classification-based approaches, content-based recommendations, and non-personalized baselines. |
