Learning to recommend from positive evidence

ABSTRACT

In recent years, many systems and approaches for recommending information, products or other objects have been developed. In these systems, often machine learning methods that need training input to acquire a user interest profile are used. Such methods typically need positive and negative evidence of the user's interests. To obtain both kinds of evidence, many systems make users rate relevant objects explicitly. Others merely observe the user's behavior, which fairly obviously yields positive evidence; in order to be able to apply the standard learning methods, these systems mostly use heuristics that attempt to find also negative evidence in observed behavior. In this paper, we present several approaches to learning interest profiles from positive evidence only, as it is contained in observed user behavior. Thus, both the problem of interrupting the user for ratings and the problem of somewhat artificially determining negative evidence are avoided. The learning approaches were developed and tested in the context of the Web-based ELFI information system. It is in real use by more than 1000 people. We give a brief sketch of ELFI and describe the experiments we made based on ELFI usage logs to evaluate the different proposed methods.

REFERENCES

• 1: Bellman R. Adaptive Control Processes: A Guided Tour. Princeton University Press, 1961.
• 2: Daniel Billsus , Michael J. Pazzani, A hybrid user model for news story classification, Proceedings of the seventh international conference on User modeling, p.99-108, June 1999, Banff, Canada
• 3: Kruskal J. Multidimensional scaling by optimising goodness of fit to a non-metrical hypothesis. Psychometrika 1, l-27, 1964
• 4: Lieberman H. Letizia: An Agent That Assists Web Browsing. International Joint Conference on Artificial ZntelEigence, Montreal, 1995.
• 5: Lieberman H. Information Agents at MIT. KZ, 12, 3, 17-23, 1998.
• 6: Mitchell T. Instance-Based Learning. Chapter 8 of Machine Learning. McGraw-Hill, 1997
• 7: Mladenic D. Personal WebWatcher: Implementation and design. Technical Report, IJS, October 1996.
• 8: Achim Nick , Jürgen Koenemann , Elmar Schalück, ELFI: information brokering for the domain of research funding, Computer Networks and ISDN Systems, v.30 n.16-18, p.1491-1500, Sept. 30, 1998
• 9: Michael Pazzani , Daniel Billsus, Learning and Revising User Profiles: The Identification ofInteresting Web Sites, Machine Learning, v.27 n.3, p.313-331, June 1997 [doi>10.1023/A:1007369909943]
• 10: Wolfgang Pohl , Achim Nick, Machine learning and knowledge representation in the LaboUr approach to user modeling, Proceedings of the seventh international conference on User modeling, p.179-188, June 1999, Banff, Canada
• 11: Paul Resnick , Hal R. Varian, Recommender systems, Communications of the ACM, v.40 n.3, p.56-58, March 1997 [doi>10.1145/245108.245121]

INDEX TERMS

Primary Classification:
  I. Computing Methodologies
  I.2 ARTIFICIAL INTELLIGENCE

Additional Classification:
  I. Computing Methodologies
  I.6 SIMULATION AND MODELING

General Terms:
Design, Experimentation, Theory

Keywords:
adaptive recommendation interfaces, evaluation of methods, machine learning, user modeling