Personalized video recommendations with both historical and new items

Zhen Zhang, Zhongnan Huang, Guangyu Gao, Chi Harold Liu*

*Corresponding author for this work

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

Abstract

Recommender systems have been proven as an essential tool to solve the information overload problem due to the burst of Internet traffic, however traditional approaches only consider to recommend items that users have not seen before, and thus ignore the significance of those items in a user’s historical records. This is motivated by the fact that users often revisit those items they have watched before, especially for TV series. Based on this, in this paper, we introduce a new concept called “revisiting ratio”, to uniquely represent the ratio between the new and old items. We also propose a “preference model” to aid selecting the most related historical records. Finally, theoretical analysis and extensive results are supplemented to show the advantages of the proposed system.

Original languageEnglish
Title of host publicationCloud Computing - 5th International Conference, CloudComp 2014, Revised Selected Papers
EditorsRoy Xiaorong Lai, Victor C.M. Leung, Min Chen, Jiafu Wan
PublisherSpringer Verlag
Pages29-42
Number of pages14
ISBN (Electronic)9783319160498
DOIs
Publication statusPublished - 2015
Event5th International Conference on Cloud Computing, CloudComp 2014 - Guilin, China
Duration: 19 Oct 201421 Oct 2014

Publication series

NameLecture Notes of the Institute for Computer Sciences, Social-Informatics and Telecommunications Engineering, LNICST
Volume142
ISSN (Print)1867-8211

Conference

Conference5th International Conference on Cloud Computing, CloudComp 2014
Country/TerritoryChina
CityGuilin
Period19/10/1421/10/14

Keywords

  • Recommender system
  • Reference model
  • Revisiting ratio

Fingerprint

Dive into the research topics of 'Personalized video recommendations with both historical and new items'. Together they form a unique fingerprint.

Cite this