Improving cold music recommendation through hierarchical audio alignment

Hao Ding, Jia Huang, Houwei Cao, Yong Liu

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

Abstract

Collaborative filtering (CF) is the state-of-the-art approach to item recommendation. However, it can neither recommend new items with no user feedbacks, nor could it recommend "long-tail" items easily. Content-based filtering can solve both problems through content analysis. However, content-based filtering alone has a much worse performance than CF. In this paper, we fuse user feedbacks and content analysis into the probabilistic matrix factorization framework. In particular, we propose a recursive dynamic programming approach to computing item similarity matrix from item content. Item latent factors are predicted from the item similarity matrix when no usage data is available. We investigate how performances of recommendation algorithms vary on items with different popularities. Results show that our approach has better performance than the same hybrid model with naive item similarity measures and Matrix Factorization.

Original languageEnglish (US)
Title of host publicationProceedings - 2016 IEEE International Symposium on Multimedia, ISM 2016
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages77-82
Number of pages6
ISBN (Electronic)9781509045709
DOIs
StatePublished - Jan 18 2017
Event18th IEEE International Symposium on Multimedia, ISM 2016 - San Jose, United States
Duration: Dec 11 2016Dec 13 2016

Other

Other18th IEEE International Symposium on Multimedia, ISM 2016
CountryUnited States
CitySan Jose
Period12/11/1612/13/16

Fingerprint

Collaborative filtering
Factorization
Feedback
Electric fuses
Dynamic programming

ASJC Scopus subject areas

  • Artificial Intelligence
  • Computer Networks and Communications
  • Media Technology
  • Computer Science Applications

Cite this

Ding, H., Huang, J., Cao, H., & Liu, Y. (2017). Improving cold music recommendation through hierarchical audio alignment. In Proceedings - 2016 IEEE International Symposium on Multimedia, ISM 2016 (pp. 77-82). [7823590] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ISM.2016.90

Improving cold music recommendation through hierarchical audio alignment. / Ding, Hao; Huang, Jia; Cao, Houwei; Liu, Yong.

Proceedings - 2016 IEEE International Symposium on Multimedia, ISM 2016. Institute of Electrical and Electronics Engineers Inc., 2017. p. 77-82 7823590.

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

Ding, H, Huang, J, Cao, H & Liu, Y 2017, Improving cold music recommendation through hierarchical audio alignment. in Proceedings - 2016 IEEE International Symposium on Multimedia, ISM 2016., 7823590, Institute of Electrical and Electronics Engineers Inc., pp. 77-82, 18th IEEE International Symposium on Multimedia, ISM 2016, San Jose, United States, 12/11/16. https://doi.org/10.1109/ISM.2016.90
Ding H, Huang J, Cao H, Liu Y. Improving cold music recommendation through hierarchical audio alignment. In Proceedings - 2016 IEEE International Symposium on Multimedia, ISM 2016. Institute of Electrical and Electronics Engineers Inc. 2017. p. 77-82. 7823590 https://doi.org/10.1109/ISM.2016.90
Ding, Hao ; Huang, Jia ; Cao, Houwei ; Liu, Yong. / Improving cold music recommendation through hierarchical audio alignment. Proceedings - 2016 IEEE International Symposium on Multimedia, ISM 2016. Institute of Electrical and Electronics Engineers Inc., 2017. pp. 77-82
@inproceedings{880df44da3694bd5816b5d22822665c2,
title = "Improving cold music recommendation through hierarchical audio alignment",
abstract = "Collaborative filtering (CF) is the state-of-the-art approach to item recommendation. However, it can neither recommend new items with no user feedbacks, nor could it recommend {"}long-tail{"} items easily. Content-based filtering can solve both problems through content analysis. However, content-based filtering alone has a much worse performance than CF. In this paper, we fuse user feedbacks and content analysis into the probabilistic matrix factorization framework. In particular, we propose a recursive dynamic programming approach to computing item similarity matrix from item content. Item latent factors are predicted from the item similarity matrix when no usage data is available. We investigate how performances of recommendation algorithms vary on items with different popularities. Results show that our approach has better performance than the same hybrid model with naive item similarity measures and Matrix Factorization.",
author = "Hao Ding and Jia Huang and Houwei Cao and Yong Liu",
year = "2017",
month = "1",
day = "18",
doi = "10.1109/ISM.2016.90",
language = "English (US)",
pages = "77--82",
booktitle = "Proceedings - 2016 IEEE International Symposium on Multimedia, ISM 2016",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - Improving cold music recommendation through hierarchical audio alignment

AU - Ding, Hao

AU - Huang, Jia

AU - Cao, Houwei

AU - Liu, Yong

PY - 2017/1/18

Y1 - 2017/1/18

N2 - Collaborative filtering (CF) is the state-of-the-art approach to item recommendation. However, it can neither recommend new items with no user feedbacks, nor could it recommend "long-tail" items easily. Content-based filtering can solve both problems through content analysis. However, content-based filtering alone has a much worse performance than CF. In this paper, we fuse user feedbacks and content analysis into the probabilistic matrix factorization framework. In particular, we propose a recursive dynamic programming approach to computing item similarity matrix from item content. Item latent factors are predicted from the item similarity matrix when no usage data is available. We investigate how performances of recommendation algorithms vary on items with different popularities. Results show that our approach has better performance than the same hybrid model with naive item similarity measures and Matrix Factorization.

AB - Collaborative filtering (CF) is the state-of-the-art approach to item recommendation. However, it can neither recommend new items with no user feedbacks, nor could it recommend "long-tail" items easily. Content-based filtering can solve both problems through content analysis. However, content-based filtering alone has a much worse performance than CF. In this paper, we fuse user feedbacks and content analysis into the probabilistic matrix factorization framework. In particular, we propose a recursive dynamic programming approach to computing item similarity matrix from item content. Item latent factors are predicted from the item similarity matrix when no usage data is available. We investigate how performances of recommendation algorithms vary on items with different popularities. Results show that our approach has better performance than the same hybrid model with naive item similarity measures and Matrix Factorization.

UR - http://www.scopus.com/inward/record.url?scp=85015171535&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85015171535&partnerID=8YFLogxK

U2 - 10.1109/ISM.2016.90

DO - 10.1109/ISM.2016.90

M3 - Conference contribution

AN - SCOPUS:85015171535

SP - 77

EP - 82

BT - Proceedings - 2016 IEEE International Symposium on Multimedia, ISM 2016

PB - Institute of Electrical and Electronics Engineers Inc.

ER -