Feature-based and Clique-based User Models for Movie Selection: A Comparative Study

  • Authors:
  • Joshua Alspector;Aleksander Koicz;N. Karunanithi

  • Affiliations:
  • ECE Department, University of Colorado, Colorado Springs, Colorado 90918, U.S.A. e-mail: josh@eas.uccs.edu;ECE Department, University of Colorado, Colorado Springs, Colorado 90918, U.S.A. e-mail: ark@eas.uccs.edu;IF-319B, Bellcore, 445 South Street, Morristown, NJ 07960. e-mail: karun@bellcore.com

  • Venue:
  • User Modeling and User-Adapted Interaction
  • Year:
  • 1997

Quantified Score

Hi-index 0.00

Visualization

Abstract

The huge amount of information available in thecurrently evolving world wide information infrastructure at any onetime can easily overwhelm end-users. One way to address theinformation explosion is to use an ’information filtering agent‘which can select information according to the interest and/or needof an end-user. However, at present few information filtering agentsexist for the evolving world wide multimedia information infrastructure.In this study, we evaluate the use of feature-based approachesto user modeling with the purpose of creating a filtering agentfor the video-on-demand application. We evaluate severalfeature and clique-based models for 10 voluntary subjects whoprovided ratings for the movies. Our preliminary results suggestthat feature-based selection can be a useful tool torecommend movies according to the taste of the userand can be as effective as a movie rating expert. Wecompare our feature-based approach with a clique-based approach, whichhas advantages where information from other users is available.