Comparing Multi-issue Multi-lateral Negotiation Approaches for Group Recommendation

SCHIAFFINO, SILVIA; MONTESERIN, ARIEL; QUINTERO, EMANUEL

Conferencia; Mexican International Conference on Artificial Intelligence; 2020

Recommender systems help users to deal with information overload when trying to find interesting items to consume in a certain domain (e-commerce, education, entertainment, tourism). Recommendations have been traditionally made to individuals, but in the last years recommender systems have been applied to group of users. Several aggregation techniques have been proposed to generate group recommendations, but most of them present some shortcomings when trying to satisfy the whole group evenly. In addition to recommendations to multiple users, it also arises the need of recommending multiple items to each group. In this context, this work proposes a multi-agent approach that uses a multi-agent system to make recommendations to a group of users using negotiation techniques. This approach also considers the recommendation of multiple items. Two alternatives within the approach were evaluated in the movies domain with promising results.