- Mapping domain-specific recommendation tasks onto a general recommendation pipeline [k3]
- Implementation of recommendation baselines (random, popularity, demographic popularity) [k6]
- Implementation of various collaborative-filtering-based recommenders (e.g., neighbourhood-based and using matrix factorization) [k6]
- Implementation of various content-based recommenders (various user representations and scoring functions) [k6]
- Implementation of setups for evaluating accuracy and beyond-accuracy aspects of recommender systems [k6]
- Systematic performance analysis of recommender systems (data splits, hyper-parameter search, per-user evaluation) [k4,k5,k6]
|
- Key components of recommender systems and their interaction
- Conceptual nuances of explicit and implicit feedback in recommender systems
- Core motivations for evaluation of recommender systems
- Fundamentals of accuracy and beyond accuracy evaluation metrics
- Knowledge on methodologies for informed comparison of recommender systems
- Popularity calibration as a way to estimate societal impact and beyond-accuracy aspects of recommender systems
|
