An Opportunistic Music Sharing System Based on Mobility Prediction and Preference Learning

With the widespread use of smart phones, opportunistic networks that leverage opportunistic contacting and data transmission among people have become a new data sharing medium. In this paper, we propose an intelligent music sharing system called "Music On Go", which is based on mobility prediction and preference extraction in opportunistic networks. In detail, we first present a human mobility prediction method based on geo-trajectory mining. A tag-based preference learning method is then proposed to extract user preference for supporting file sharing among peers. We implemented the system in commercial smart phones equipped with Bluetooth and GPS interfaces. Experimental results showed that the proposed system successfully enables opportunistic mobile file sharing and facilitates user interactions.