Not Logged In

Hierarchical Probabilistic Relational Models for Collaborative Filtering

Full Text: PRM-CF.pdf PDF
Other Attachments: HierarchicalPRMsForCF.ppt [Poster] PPT

This paper applies Probabilistic Relational Models (PRMs) to the Collaborative Filtering task, focussing on the EachMovie data set. We first learn a standard PRM, and show that its performance is competitive with the best known techniques. We then define a hierarchical PRM, which extends standard PRMs by dynamically refining classes into hierarchies, which improves the expressiveness as well as the context sensitivity of the PRM. Finally, we show that hierarchical PRMs achieve state-of-the-art results on this dataset.

Citation

J. Newton, R. Greiner. "Hierarchical Probabilistic Relational Models for Collaborative Filtering". SRL2004: Statistical Relational Learning and its Connections to Other Fields, July 2004.

Keywords: probabilistic relational models, Belief Networks, machine learning
Category: In Workshop

BibTeX

@misc{Newton+Greiner:SRL200404,
  author = {Jack Newton and Russ Greiner},
  title = {Hierarchical Probabilistic Relational Models for Collaborative
    Filtering},
  booktitle = {SRL2004: Statistical Relational Learning and its Connections to
    Other Fields},
  year = 2004,
}

Last Updated: June 06, 2007
Submitted by Russ Greiner

University of Alberta Logo AICML Logo