A predictive model of music preference using pairwise comparisons

Bjørn Sand Jensen; Javier Saez  Gallego; Jan Larsen

doi:10.1109/ICASSP.2012.6288294

A predictive model of music preference using pairwise comparisons

Bjørn Sand Jensen, Javier Saez Gallego, Jan Larsen

Technical University of Denmark

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

Abstract

Music recommendation is an important aspect of many streaming services and multi-media systems, however, it is typically based on so-called collaborative filtering methods. In this paper we consider the recommendation task from a personal viewpoint and examine to which degree music preference can be elicited and predicted using simple and robust queries such as pairwise comparisons. We propose to model - and in turn predict - the pairwise music preference using a very flexible model based on Gaussian Process priors for which we describe the required inference. We further propose a specific covariance function and evaluate the predictive performance on a novel dataset. In a recommendation style setting we obtain a leave-one-out accuracy of 74% compared to 50% with random predictions, showing potential for further refinement and evaluation.

Original language	English
Title of host publication	2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Publisher	IEEE
Publication date	2012
Pages	1977-1980
ISBN (Print)	978-1-4673-0045-2
ISBN (Electronic)	978-1-4673-0044-5
DOIs	https://doi.org/10.1109/ICASSP.2012.6288294
Publication status	Published - 2012
Event	2012 IEEE International Conference on Acoustics, Speech and Signal Processing - Kyoto International Conference Centre, Kyoto, Japan Duration: 25 Mar 2012 → 30 Mar 2012 Conference number: 37 http://www.icassp2012.com/

Conference

Conference	2012 IEEE International Conference on Acoustics, Speech and Signal Processing
Number	37
Location	Kyoto International Conference Centre
Country/Territory	Japan
City	Kyoto
Period	25/03/2012 → 30/03/2012
Internet address	http://www.icassp2012.com/

Series	I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings
ISSN	1520-6149

Keywords

Music Preference
Kernel Methods
Gaussian Process Priors
Recommendation

Access to Document

10.1109/ICASSP.2012.6288294

OpenUrl availability

Full text

Cite this

@inproceedings{f501c428218d4132ba0cfaa9eafb547c,

title = "A predictive model of music preference using pairwise comparisons",

abstract = "Music recommendation is an important aspect of many streaming services and multi-media systems, however, it is typically based on so-called collaborative filtering methods. In this paper we consider the recommendation task from a personal viewpoint and examine to which degree music preference can be elicited and predicted using simple and robust queries such as pairwise comparisons. We propose to model - and in turn predict - the pairwise music preference using a very flexible model based on Gaussian Process priors for which we describe the required inference. We further propose a specific covariance function and evaluate the predictive performance on a novel dataset. In a recommendation style setting we obtain a leave-one-out accuracy of 74% compared to 50% with random predictions, showing potential for further refinement and evaluation.",

keywords = "Music Preference, Kernel Methods, Gaussian Process Priors, Recommendation",

author = "Jensen, {Bj{\o}rn Sand} and Gallego, {Javier Saez} and Jan Larsen",

year = "2012",

doi = "10.1109/ICASSP.2012.6288294",

language = "English",

isbn = "978-1-4673-0045-2",

series = "I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings",

publisher = "IEEE",

pages = "1977--1980 ",

booktitle = "2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)",

address = "United States",

note = "2012 IEEE International Conference on Acoustics, Speech and Signal Processing, ICASSP 2012 ; Conference date: 25-03-2012 Through 30-03-2012",

url = "http://www.icassp2012.com/",

}

Jensen, BS, Gallego, JS & Larsen, J 2012, A predictive model of music preference using pairwise comparisons. in 2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings, pp. 1977-1980 , 2012 IEEE International Conference on Acoustics, Speech and Signal Processing, Kyoto, Japan, 25/03/2012. https://doi.org/10.1109/ICASSP.2012.6288294

A predictive model of music preference using pairwise comparisons. / Jensen, Bjørn Sand; Gallego, Javier Saez ; Larsen, Jan.
2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP). IEEE, 2012. p. 1977-1980 (I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Article in proceedings › Research › peer-review

TY - GEN

T1 - A predictive model of music preference using pairwise comparisons

AU - Jensen, Bjørn Sand

AU - Gallego, Javier Saez

AU - Larsen, Jan

N1 - Conference code: 37

PY - 2012

Y1 - 2012

N2 - Music recommendation is an important aspect of many streaming services and multi-media systems, however, it is typically based on so-called collaborative filtering methods. In this paper we consider the recommendation task from a personal viewpoint and examine to which degree music preference can be elicited and predicted using simple and robust queries such as pairwise comparisons. We propose to model - and in turn predict - the pairwise music preference using a very flexible model based on Gaussian Process priors for which we describe the required inference. We further propose a specific covariance function and evaluate the predictive performance on a novel dataset. In a recommendation style setting we obtain a leave-one-out accuracy of 74% compared to 50% with random predictions, showing potential for further refinement and evaluation.

AB - Music recommendation is an important aspect of many streaming services and multi-media systems, however, it is typically based on so-called collaborative filtering methods. In this paper we consider the recommendation task from a personal viewpoint and examine to which degree music preference can be elicited and predicted using simple and robust queries such as pairwise comparisons. We propose to model - and in turn predict - the pairwise music preference using a very flexible model based on Gaussian Process priors for which we describe the required inference. We further propose a specific covariance function and evaluate the predictive performance on a novel dataset. In a recommendation style setting we obtain a leave-one-out accuracy of 74% compared to 50% with random predictions, showing potential for further refinement and evaluation.

KW - Music Preference

KW - Kernel Methods

KW - Gaussian Process Priors

KW - Recommendation

U2 - 10.1109/ICASSP.2012.6288294

DO - 10.1109/ICASSP.2012.6288294

M3 - Article in proceedings

SN - 978-1-4673-0045-2

T3 - I E E E International Conference on Acoustics, Speech and Signal Processing. Proceedings

SP - 1977

EP - 1980

BT - 2012 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)

PB - IEEE

T2 - 2012 IEEE International Conference on Acoustics, Speech and Signal Processing

Y2 - 25 March 2012 through 30 March 2012

ER -

A predictive model of music preference using pairwise comparisons

Abstract

Conference

Keywords

Access to Document

OpenUrl availability

Fingerprint

Cite this