Résumé
To simplify effective music filtering, recommender systems (RS) have received great attention from both industry and academia area. To select which music to recommend, traditional RS uses an approximation of users’ real interests. However, while discarding users’ contexts, profiles information is not able to reflect their exact needs and to provide overpowering recommendations. One of the main issues that have to be considered before the conception of context-aware recommender systems (CARS) is the estimation of the relevance of contextual information. The use of irrelevant or superfluous contextual factors can generate serious problems about the complexity and the quality of recommendations. In this paper, we introduce a multi-dimensional context model for music CARS. We started by the acquisition of explicit items rating from a population in various possible contextual situations. Thus, we proposed a user-based methodology aiming to judge the relation between contextual factors and musical genres. Next, we applied the Multiple Linear Regression technique on users’ perceived ratings, to define an order of importance between contextual dimensions. We described raw collected data with basic statistics about the created dataset. We also summarized the key results and discussed key findings. Finally, we propose a new framework for Music CARS.