Preference Elicitation and Negotiation in a Group Recommender System

Abstract

Abstract. We present a novel approach to group recommender systems that better takes into account the social interaction in a group when formulating, discussing and negotiating the features of the item to be jointly selected. Our approach provides discussion support in a collaborative preference elicitation and negotiation process. Individual preferences are continuously aggregated and immediate feedback of the resulting recommendations is provided. We also support the last stage in the decision process when users collectively select the final item from the recommendation set. The prototype hotel recommender Hootle is developed following these concepts and tested in a user study. The results indicate a higher overall satisfaction with the system as well as a higher perceived recommendation quality when compared against a system version where no negotiation was possible. However, they also indicate that the negotiation-based approach may be more suitable for smaller groups, an aspect that will require further research.

Negotiation and Reconciliation of Preferences in a Group Recommender System

“…Following this idea, our approach allows a group of users to collaboratively create and discuss a preference model (thus addressing collaborative [34] and explicit [33] preference elicitation). A first prototype was designed [1] where users were able to create their own individual lists of features ordered by importance, obtaining immediate feedback on the aggregated group's preference model and its matching recommended items. The results obtained from the consequent user evaluation were promising, suggesting that our approach effectively improves the quality of recommendations when compared against standard group recommender systems.…”

“…These findings led to the development of a third prototype, Hootle Mobile, based on a revised, streamlined method where private spaces have been completely removed and preferences could be directly added to the group model. This paper provides an aggregated and extended account of work reported in a prior publication [2], incorporating a design synopsis and empirical findings from a first version of the system [1] as well as more details on its empirical evaluation. In addition, we report for the first time on a mobile version of the system that also modifies the approach by directly expressing user preferences in the shared group space.…”