A Destination Recommendation System for Virtual Worlds

This paper describes the design and training of a recommendation system that helps the user navigate in Second Life (SL), a massively multi-player online environment. Second Life allows users to create a virtual avatar and explore areas built by other users. However, unlike the real world, virtual attractions can be constructed within hours and previous ones often rapidly fall into disuse. Without recent information about the state of regions in the world, it is difficult to assist the user’s searches in the virtual world. In this paper, we present a framework for leveraging recommendations about locations made by other users using a collaborative item-based filtering method in which locations are treated as items. Our system runs in real-time and presents the user with information via a virtual heads-up display (HUD).