Hybrid collaborative and content-based music recommendation using probabilistic model with latent user preferences

Kazuyoshi Yoshii, Masataka Goto, Kazunori Komatani, Tetsuya Ogata, Hiroshi G. Okuno

Research output: Chapter in Book/Report/Conference proceedingConference contribution

82 Citations (Scopus)

Abstract

This paper presents a hybrid music recommendation method that solves problems of two prominent conventional methods: collaborative filtering and content-based recommendation. The former cannot recommend musical pieces that have no ratings because recommendations are based on actual user ratings. In addition, artist variety in recommended pieces tends to be poor. The latter, which recommends musical pieces that are similar to users' favorites in terms of music content, has not been fully investigated. This induces unreliability in modeling of user preferences; the content similarity does not completely reflect the preferences. Our method integrates both rating and content data by using a Bayesian network called an aspect model. Unobservable user preferences are directly represented by introducing latent variables, which are statistically estimated. To verify our method, we conducted experiments by using actual audio signals of Japanese songs and the corresponding rating data collected from Amazon. The results showed that our method outperforms the two conventional methods in terms of recommendation accuracy and artist variety and can reasonably recommend pieces even if they have no ratings.

Original languageEnglish
Title of host publicationISMIR 2006 - 7th International Conference on Music Information Retrieval
Pages296-301
Number of pages6
Publication statusPublished - 2006
Externally publishedYes
Event7th International Conference on Music Information Retrieval, ISMIR 2006 - Victoria, BC
Duration: 2006 Oct 82006 Oct 12

Other

Other7th International Conference on Music Information Retrieval, ISMIR 2006
CityVictoria, BC
Period06/10/806/10/12

Fingerprint

Collaborative filtering
Bayesian networks
Experiments
Statistical Models
Music
User Preferences
Rating

Keywords

  • Collaborative filtering
  • Content-based recommendation
  • Hybrid method
  • Probabilistic model

ASJC Scopus subject areas

  • Music
  • Information Systems

Cite this

Yoshii, K., Goto, M., Komatani, K., Ogata, T., & Okuno, H. G. (2006). Hybrid collaborative and content-based music recommendation using probabilistic model with latent user preferences. In ISMIR 2006 - 7th International Conference on Music Information Retrieval (pp. 296-301)

Hybrid collaborative and content-based music recommendation using probabilistic model with latent user preferences. / Yoshii, Kazuyoshi; Goto, Masataka; Komatani, Kazunori; Ogata, Tetsuya; Okuno, Hiroshi G.

ISMIR 2006 - 7th International Conference on Music Information Retrieval. 2006. p. 296-301.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Yoshii, K, Goto, M, Komatani, K, Ogata, T & Okuno, HG 2006, Hybrid collaborative and content-based music recommendation using probabilistic model with latent user preferences. in ISMIR 2006 - 7th International Conference on Music Information Retrieval. pp. 296-301, 7th International Conference on Music Information Retrieval, ISMIR 2006, Victoria, BC, 06/10/8.
Yoshii K, Goto M, Komatani K, Ogata T, Okuno HG. Hybrid collaborative and content-based music recommendation using probabilistic model with latent user preferences. In ISMIR 2006 - 7th International Conference on Music Information Retrieval. 2006. p. 296-301
Yoshii, Kazuyoshi ; Goto, Masataka ; Komatani, Kazunori ; Ogata, Tetsuya ; Okuno, Hiroshi G. / Hybrid collaborative and content-based music recommendation using probabilistic model with latent user preferences. ISMIR 2006 - 7th International Conference on Music Information Retrieval. 2006. pp. 296-301
@inproceedings{c9263609497442e1b98d12ded7f62a98,
title = "Hybrid collaborative and content-based music recommendation using probabilistic model with latent user preferences",
abstract = "This paper presents a hybrid music recommendation method that solves problems of two prominent conventional methods: collaborative filtering and content-based recommendation. The former cannot recommend musical pieces that have no ratings because recommendations are based on actual user ratings. In addition, artist variety in recommended pieces tends to be poor. The latter, which recommends musical pieces that are similar to users' favorites in terms of music content, has not been fully investigated. This induces unreliability in modeling of user preferences; the content similarity does not completely reflect the preferences. Our method integrates both rating and content data by using a Bayesian network called an aspect model. Unobservable user preferences are directly represented by introducing latent variables, which are statistically estimated. To verify our method, we conducted experiments by using actual audio signals of Japanese songs and the corresponding rating data collected from Amazon. The results showed that our method outperforms the two conventional methods in terms of recommendation accuracy and artist variety and can reasonably recommend pieces even if they have no ratings.",
keywords = "Collaborative filtering, Content-based recommendation, Hybrid method, Probabilistic model",
author = "Kazuyoshi Yoshii and Masataka Goto and Kazunori Komatani and Tetsuya Ogata and Okuno, {Hiroshi G.}",
year = "2006",
language = "English",
isbn = "9781550583496",
pages = "296--301",
booktitle = "ISMIR 2006 - 7th International Conference on Music Information Retrieval",

}

TY - GEN

T1 - Hybrid collaborative and content-based music recommendation using probabilistic model with latent user preferences

AU - Yoshii, Kazuyoshi

AU - Goto, Masataka

AU - Komatani, Kazunori

AU - Ogata, Tetsuya

AU - Okuno, Hiroshi G.

PY - 2006

Y1 - 2006

N2 - This paper presents a hybrid music recommendation method that solves problems of two prominent conventional methods: collaborative filtering and content-based recommendation. The former cannot recommend musical pieces that have no ratings because recommendations are based on actual user ratings. In addition, artist variety in recommended pieces tends to be poor. The latter, which recommends musical pieces that are similar to users' favorites in terms of music content, has not been fully investigated. This induces unreliability in modeling of user preferences; the content similarity does not completely reflect the preferences. Our method integrates both rating and content data by using a Bayesian network called an aspect model. Unobservable user preferences are directly represented by introducing latent variables, which are statistically estimated. To verify our method, we conducted experiments by using actual audio signals of Japanese songs and the corresponding rating data collected from Amazon. The results showed that our method outperforms the two conventional methods in terms of recommendation accuracy and artist variety and can reasonably recommend pieces even if they have no ratings.

AB - This paper presents a hybrid music recommendation method that solves problems of two prominent conventional methods: collaborative filtering and content-based recommendation. The former cannot recommend musical pieces that have no ratings because recommendations are based on actual user ratings. In addition, artist variety in recommended pieces tends to be poor. The latter, which recommends musical pieces that are similar to users' favorites in terms of music content, has not been fully investigated. This induces unreliability in modeling of user preferences; the content similarity does not completely reflect the preferences. Our method integrates both rating and content data by using a Bayesian network called an aspect model. Unobservable user preferences are directly represented by introducing latent variables, which are statistically estimated. To verify our method, we conducted experiments by using actual audio signals of Japanese songs and the corresponding rating data collected from Amazon. The results showed that our method outperforms the two conventional methods in terms of recommendation accuracy and artist variety and can reasonably recommend pieces even if they have no ratings.

KW - Collaborative filtering

KW - Content-based recommendation

KW - Hybrid method

KW - Probabilistic model

UR - http://www.scopus.com/inward/record.url?scp=84873459337&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84873459337&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9781550583496

SP - 296

EP - 301

BT - ISMIR 2006 - 7th International Conference on Music Information Retrieval

ER -