SpIteR: a Module for Recommending Dynamic Personalized Museum Tours Pierpaolo Basile, Marco de Gemmis, Leo Iaquinta, Pasquale Lops, Cataldo Musto, Fedelucio Narducci, Giovanni Semeraro Department of Computer Science, University “Aldo Moro” Bari, Italy {basilepp, degemmis, iaquinta, lops, musto, narducci, semeraro}@di.uniba.it Abstract—Recommender systems (RSs) proved to make easier the task of accessing relevant information in a broad range of domains. In content-based RSs, preferences on content items expressed by users turned out to be reliable indicators to suggest and filter interesting contents. Item representation plays a key role in content-based RSs, thus choosing proper facets to represent items is a fundamental task for deploying effective RSs. Contextual facets are often marginally relevant to predict user preferences, but in some domains disregarding contextual facets makes recommendations useless. This paper proposes a strategy to improve the effectiveness of a content-based RS that dynamically suggests tours within a museum by exploiting contextual facets such the physical layout of items and the interaction of users with the environment. Keywords-Context-aware Recommender Systems; Cultural Heritage Fruition; User-generated Contents I. I NTRODUCTION The importance of providing digital access to cultural heritage collections has been already acknowledged by mu- seums for almost four decades [1]. More recently, it has been asserted that the personalization drives “the museum monologue” into “a user-centred information dialog” be- tween the museum and its visitors [2]. Visitors spend more and more time to discover interesting artworks, prepare a museum tour, or learn related knowledge about artworks, usually in relation to a (possible) physical museum visit [3]. Cultural heritage personalization refers to supporting visitors in the selection and filtering of preferred artifacts, and in the creation of personalized tours. Since RSs have proved to be useful in helping users access to desired information (especially in domains where users are not expert or familiar with) [4], they have found their way also in the context of museums. The goal of a RS is finding out the k most interesting items for the current user. Thus, prominent aspects of users and items are scanned and similarity techniques are employed to determine the most relevant suggestions. Among techniques used to design RSs, the most investigated ones concern content or metadata of items in the content-based approach, and the user’s social environment in the collaborative filtering approach. The user active role with respect to RSs concerns tra- ditionally only to provide a relevance feedback (rating) on items and suggestions. The Web 2.0 (r)evolution is encourag- ing people to overstep the passive role of content consumers. Thus, the user increased participation can be exploited in recommendation process. One of the forms of user-generated content (UGC) that has drawn more attention is tagging, i.e. resource annotation with free keywords, that builds a socially-constructed classification schema, called folksonomy (folks + taxonomy). Social tagging in the cultural heritage personalization scenario promotes the audience engagement with museum collections. Providing access based on the resulting folksonomy opens museum collections to new interpretations, which reflect also visitor perspectives and not only curator ones, and helps to bridge the gap between the professional language of the curator and the popular language of the museum visitor. Beside encouraging the user increased participation, in order to make user experience enthralling, the item arrange- ment as well the user interaction with the environment are very relevant. Indeed, a static ordered list in accordance with only the assessed interests is not the best strategy to present the most interesting items, e.g. tortuous paths with repetitive passages make the user disoriented especially under a time constraint. Finally, different users interact with the environment in different manner, e.g. they travel with different speed, they spend different time to admire artworks, they divert from suggested tour. This paper presents the SpIteR (Spatial Item Recom- mender) module that suggests dynamic context-sensitive tours according to user profiles learned on static contents and UGCs. Indeed, the module uses contextual facets to adjust (“contextualize”) the resulting set of content-based recommendations in post-filtering [5]. In the “virtuous cir- cle” vision [6], i.e. the connection between on-line and on- site experiences, SpIteR is a preliminary attempt to react to the user behavior contextualizing the personalized tour that is an open issues in cultural heritage fruition, for instance for the CHIP project as recently reported by Wang et al. [3]. The remainder of this paper is structured as follows. Section II describes the our recommender system. Section III discusses the case study. Finally, Section IV draws conclu- sions and provides direction for future work. II. ARCHITECTURE The recommendation process is performed in several steps, each of which is handled by a separate component of 2009 IEEE/WIC/ACM International Joint Conference on Web Intelligence and Intelligent Agent Technology 978-0-7695-3801-3/09 $26.00 © 2009 IEEE DOI 10.1109/WI-IAT.2009.99 584