Collaborative ranking

Suhrid Balakrishnan, Sumit Chopra

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

Abstract

Typical recommender systems use the root mean squared error (RMSE) between the predicted and actual ratings as the evaluation metric. We argue that RMSE is not an optimal choice for this task, especially when we will only recommend a few (top) items to any user. Instead, we propose using a ranking metric, namely normalized discounted cumulative gain (NDCG), as a better evaluation metric for this task. Borrowing ideas from the learning to rank community for web search, we propose novel models which approximately optimize NDCG for the recommendation task. Our models are essentially variations on matrix factorization models where we also additionally learn the features associated with the users and the items for the ranking task. Experimental results on a number of standard collaborative filtering data sets validate our claims. The results also show the accuracy and efficiency of our models and the benefits of learning features for ranking.

Original languageEnglish (US)
Title of host publicationWSDM 2012 - Proceedings of the 5th ACM International Conference on Web Search and Data Mining
Pages143-152
Number of pages10
DOIs
StatePublished - 2012
Event5th ACM International Conference on Web Search and Data Mining, WSDM 2012 - Seattle, WA, United States
Duration: Feb 8 2012Feb 12 2012

Publication series

NameWSDM 2012 - Proceedings of the 5th ACM International Conference on Web Search and Data Mining

Other

Other5th ACM International Conference on Web Search and Data Mining, WSDM 2012
Country/TerritoryUnited States
CitySeattle, WA
Period2/8/122/12/12

Keywords

  • Collaborative ranking
  • Learning to rank
  • NDCG
  • Recommender systems
  • RMSE

ASJC Scopus subject areas

  • Computer Networks and Communications

Fingerprint

Dive into the research topics of 'Collaborative ranking'. Together they form a unique fingerprint.

Cite this