Research Seminars

The Long Tail or the Long Fail of Music Recommendation?

Oscar Celma
Music Technology Group (MTG), Pompeu Fabra University, Barcelona (Spain)

Friday 30 January 2009, 15:00, Room 105

Abstract

Music consumption is biased towards a few popular artists. For instance, in 2007 only 1% of all digital tracks accounted for 80% of all sales. Similarly, 1,000 albums accounted for 50% of all album sales, and 80% of all albums sold were purchased less than 100 times. There is a need to assist people to filter, discover, personalise and recommend from the huge amount of music content available along the Long Tail.

Current music recommendation algorithms try to accurately predict what people demand to listen to. However, quite often these algorithms tend to recommend popular (or well-known to the user) music, decreasing the effectiveness of the recommendations. These approaches focus on improving the accuracy of the recommendations. That is, try to make accurate predictions about what a user could listen to, or buy next, independently of how useful to the user the provided recommendations could be. Yet, effective recommendation systems should also promote novel and relevant material (non-obvious recommendations), taken primarily from the tail of the music popularity distribution.

In this seminar we highlight the main differences of three music recommendation approaches: social-based (using last.fm data), audio content-based, and a hybrid approach that combines content-based and human experts. To evaluate the novelty and effectiveness factors of the three approaches, we apply complex network metrics and we combine them with the information about item popularity. Furthermore, to asses the user's relevance and novelty of the recommendations, we present the results of a user-based survey that compares the three recommendation approaches.

Biography

Oscar Celma is a researcher at Music Technology Group (MTG) since 2000, and Lecturer at the Pompeu Fabra University, Barcelona (Spain). He is also a co-founder of the BMAT company, a spin-off of the MTG. Since 2006 he is an Invited Expert of the W3C Multimedia Semantics Incubator Group.

The main focus of his research lies in the music recommendation field. In 2006, Oscar received the 2nd prize in the International Semantic Web Challenge for the system named Foafing the Music, a personalized music recommendation that exploits music related information available from the web.

During his undergraduate studies, he also obtained the diplomas in classical guitar, and composition. Though, nowadays he only makes some noise with his old Grestch.

The list of publications of the speaker is available at http://mtg.upf.edu/biblio/author/Celma