User Preferences in Tourist Itineraries Recommendation Pierpaolo Di Bitonto, Francesco Di Tria, Maria Laterza, Teresa Roselli, Veronica Rossano, and Filippo Tangorra Department of Computer Science - University of Bari “Aldo Moro” Via Orabona 4, 70125 – Bari, Italy {dibitonto, francescoditria, marialaterza, roselli, rossano, tangorra}@di.uniba.it Abstract One of the most interesting current challenges in the e-tourism field is to offer services that are able to suggest attractions or event itineraries that fit tourist’s needs and preferences. In this scenario, we defined a method for generating itineraries of intangible cultural heritage events (processions, special markets, festivals, and so on) based on a theoretical model and the transitive closure computation. The method was, then, implemented, in a knowledge-based recommender system prototype, named T-Path, which is able to suggest event itineraries in Apulia region (in southern Italy). In this paper we propose an approach to identify the itineraries that best fit the tourist/user preferences. At this aim, first of all the type of each itinerary is discovered, and then the itineraries are ordered on the basis of their adequacy to the tourist preferences. Keywords: Recommender system; Tourist itineraries; Cultural heritage events 1 Introduction In the latest years the tourism domain has shown a progressive change. More and more people plan their holidays on Internet, surfing on the web in order to search new destinations, flights, accommodations, and so on. Moreover, during their stay the tourists continuously access to the web for searching interesting attraction or events. The current trend in the tourism domain is not only to build systems that are able to suggest and supply information about tourist spots such as temples, churches, museums (Scherp & Boll, 2004), or restaurants (Ricci, & Nguyen, 2004; Yap, Tan, & Pang, 2007), but also to design and implement tourist guides that are able to suggest interesting itineraries to the user (Kramer et al., 2006; Biuk-Aghai et al., 2008; De Choudhury et al., 2010) like a human guide. The main contribution of this research in the tourism field is the proposal of a recommendation process that, indentifying the itineraries that best fit the tourist preferences, supports the educational aspect of the tourism. In other words, the process proposed is able to suggest a set of event that is interesting for the user from a cultural point of view. In this context, we defined a method for automated generation of itineraries of intangible cultural heritage events (such as festivals, special markets, processions, etc.) in a knowledge-based recommender system. The method defined is based on a theoretical model (Di Bitonto et al., 2010a) that consists of a set of functions to characterize each event and a space- time relation. It aims to find chains of events using transitive closure computation. Moreover, the method was implemented in a knowledge-based recommender prototype, named T-Path (Di Bitonto et al., 2010b), that is able to build and suggest