As collections of digital music become larger and more widespread, there is a growing need for assistance in a user's navigation and interaction with a collection and with the individual members of that collection. Examining pairwise song relationships and similarities, based upon content derived features, provides a useful tool to do so. This paper looks into a means of extending a song classification algorithm to provide song to song similarity information. In order to evaluate the effectiveness of this method, the similarity data is used to group the songs into k-means clusters, these clusters are then compared against the original genre sorting algorithm.
Click to purchase paper as a non-member or login as an AES member. If your company or school subscribes to the E-Library then switch to the institutional version. If you are not an AES member and would like to subscribe to the E-Library then Join the AES!
This paper costs $33 for non-members and is free for AES members and E-Library subscribers.