Apollo - A Hybrid Recommender for Museums and Cultural Tourism

摘要

This paper introduces Apollo, a novel hybrid recommender for free-roaming or guided museum visits and cultural tourism. The recommender is based on a new conceptualisation of a visit and the adoption of a minimax (or `conservative') approach towards user satisfaction modelling. The approach is based on the integration of temporal, spatial and content dynamics captured during a visit and contribute to an estimate of the user satisfaction and a development of an optimal route as a sequence of points of interest. Apollo follows a minimax approach by targeting the minimisation of the highest possible user dissatisfaction or disengagement, rather than looking for a maximisation of the user satisfaction. A considerable effort has been dedicated to the creation of realistic simulation data for items (the exhibits), users (the visitors), and a small amount of ratings of the items by some of the users with specific characteristics selected to represent a realistic scenario. Extensive visit simulations have been conducted and results show a considerable decrease of the probable user dissatisfaction in relation to a baseline recommender.