Research and realization of music recommender algorithm based on hybrid collaborative filtering

Haiying Che, Zishi Wang

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Digital music has penetrated every aspect of people’s lives and massive resources of online digital music have maintained a rapid growth rate because of the leap of Internet. The traditional music information retrieval systems, therefore, could not meet the growing needs of users. This paper combines SVD with memory-based collaborative filtering, compensates the negative effects of sparse matrices, and then forms a hybrid recommendation algorithm. Experimental results in “R3 Yahoo!Music” shows the advantage of our proposal in accuracy.

Original languageEnglish
Title of host publicationSignal and Information Processing, Networking and Computers - Proceedings of the 1st International Congress on Signal and Information Processing, Networking and Computers, ICSINC 2015
EditorsNa Chen, Tingting Huang
PublisherCRC Press/Balkema
Pages195-202
Number of pages8
ISBN (Print)9781138028814
DOIs
Publication statusPublished - 2016
Externally publishedYes
Event1st International Congress on Signal and Information Processing, Networking and Computers, ICSINC 2015 - Beijing, China
Duration: 17 Oct 201618 Oct 2016

Publication series

NameSignal and Information Processing, Networking and Computers - Proceedings of the 1st International Congress on Signal and Information Processing, Networking and Computers, ICSINC 2015

Conference

Conference1st International Congress on Signal and Information Processing, Networking and Computers, ICSINC 2015
Country/TerritoryChina
CityBeijing
Period17/10/1618/10/16

Keywords

  • Hybrid collaborative filtering
  • KNN-based collaborative filtering
  • Music recommender system
  • SVD

Fingerprint

Dive into the research topics of 'Research and realization of music recommender algorithm based on hybrid collaborative filtering'. Together they form a unique fingerprint.

Cite this