Personalized recommender system for event attendees

Abstract

This paper presents a smart event mobile application prototype featuring beacon-based indoor geolocation, personalized recommendations and automated user interface adaptation. Focus is given here to the recommendation function, which exploits professional social network mining and semantic concept identification. We report its use in real conditions during a conference, discuss the usefulness of recommendations and more precisely the impact of the recommendation features. We can show that through appropriate semantic enrichment the conversion rate increased by factor two and the system was able to introduce previously unseen/unknown items to the users thus enhancing serendipity.