An Approach to Automatically Tracking Music Preference on Mobile Players

More and more music is being made available to the music listener today, while people have their favorite music on their mobile players. In this paper, we investigate an approach to automatically updating the music on the mobile player based on personal listening behavior. The aim is to automatically discard those pieces of music from the player the listener is fed up with, while new music is automatically selected from a large amount of available music. The source of new music could be a flat rate music delivery service, where the user pays a monthly fee to have access to a large amount of music. We assume a scenario where only a “skip” button is available to the user, which she presses when the currently playing track does not please her. We evaluate several algorithms and show that the best ones clearly outperform those with lower performance, while it remains open how much they can be improved further.