Report Number: CS-TR-98-1605
Institution: Stanford University, Department of Computer Science
Title: Learning to Surf: Multiagent Systems for Adaptive Web Page Recommendation
Author: Balabanovic, Marko
Date: March 1998
Abstract: Imagine a newspaper personalized for your tastes. Instead of a selection of articles chosen for a general audience by a human editor, a software agent picks items just for you, covering your particular topics of interest. Since there are no journalists at its disposal, the agent searches the Web for appropriate Over time, it uses your feedback on recommended articles to build a model of your interests. This thesis investigates the design of "recommender systems" which create such personalized Two research issues motivate this work and distinguish it from approaches usually taken by information retrieval or machine researchers. First, a recommender system will have many users, overlapping interests. How can this be exploited? Second, each edition of a personalized newspaper consists of a small set of articles. Techniques for deciding on the relevance of individual articles are well known, but how is the composition of the set One of the primary contributions of this research is an architecture linking populations of adaptive software agents. interests among its users are used both to increase efficiency scalability, and to improve the quality of recommendations. A interface infers document preferences by monitoring user drag-and-drop actions, and affords control over the composition sets of recommendations. Results are presented from a variety of experiments: user tests measuring learning performance, studies isolating particular tradeoffs, and usability tests investigating interaction designs.