2Laboratory for Intelligent Systems in Tourism

Mobile recommender systems have the potential to substantially enrich tourist experiences. As their handling marks a big challenge for ordinary users, its acceptance can only be evaluated when utilized by the intended user group itself- real tourists. In a field trial in Görlitz (Germany), 421 tourists explored the city with one of two different mobile information systems, a proactive recommender of personalized tours and a pull service presenting context-based information on demand. A third group of tourists was tracked by GPS receivers during their exploration of the destination relying on traditional means of information. Results point out that both mobile applications gained a high level of acceptance by providing an experience very similar to a traditional guided tour. Compared to the group tracked by GPS loggers, tourists using a mobile information system discovered four times more sights and stayed at them twice as long. Related work1 Mobile tour guides are the result of years of research in the areas of recommenders, ambient intelligence and pervasive computing. Cyberguide [Abowd, 1997]0 was one of the first mobile tour guides. Personal preferences are not analyzed to compute a tour plan but the user can retrieve information or request a route to a desired Point of Interest (PoI). GUIDE [Cheverst, 2000] is a mobile tour guide very similar to the hereby presented Dynamic Tour Guide (DTG) [Kramer, 2005]. The visitor chooses attractions from various categories. These attractions are then sequenced taking into account the opening hours, best time to visit and the distance between attractions. The sequence can be modified manually. Navigation is achieved by a map with a list of instructions. Differences to the DTG are the use of cell based positioning instead of GPS and th