Traditional recommender systems mainly focus on the accuracy of recommendation, which lead to recommender systems reinforcing popular items and ignoring lesser-known items. There is increasing evidence that providing good recommendations of surprising items can lead to better user satisfaction. Users may be delightfully surprised if long-tail items are brought to them. Marketplaces need to keep providers satisfied by making sure that their items get enough exposure. In this work, we propose a fairness-aware multi-stakeholder recommender system that uses a multi-objective evolutionary algorithm to make a trade-off between provider coverage, long-tail inclusion, personalized diversity, and recommendation accuracy. Experimental results against real-world datasets show that the proposed method significantly improves the diversity of recommended items in a personalized matter and the coverage of providers with no or minor loss of accuracy.
- Long-tail recommendation
- Multi-objective evolutionary optimization
- Multi-stakeholder recommender systems
- Personalized diversity