Music Recommender system for Wi-Fi Walkman

The research described in this report is a part of research in UseT work package of the CACTUS project. This research focuses on personalization in ad-hoc wireless environments. As a concrete example, we look into music recommender systems (MRS). In this report, we give the problem description of MRS as well as its application with the scenario of Wi-Fi walkman of the CACTUS storyboard. In this report, first, we discuss the background and motivation of our personalization research. Then we state that a recommender system (RS) is a target form to generate personalization in our particular ad-hoc wireless network domain. Second, we discuss our concrete application: Music recommendation with the Wi-Fi walkman scenario. The definition and explanation of our problem are given. Third, we conduct a literature review about RS’s by looking at commercially available products and the different research aspects. Fourth, our scope and research issues of MRS are discussed. We state that user interest should be modelled both in terms of long-term interest and short-term interest in order to achieve an accurate learning result. Finally, we conclude with two potential directions: 1). A unified model that combines recommendation data from both content and social aspects. 2). Incorporating context information in order to learn the short-term interest.