Recommender Systems

Personalized Assistants That Help Users Select Products and Services

AITopics > Applications / Expert Systems > AI Applications Areas > Recommender Systems

Definition of the Field

A recommender system is "any system that produces individualized recommendations as output or has the effect of guiding the user in a personalized way to interesting or useful objects in a large space of possible options."
From Burke, R. Hybrid Recommender Systems: Survey and Experiments. User Modeling and User-Adapted Interaction. 12(4), pages 331-370.

Overview

Selecting and Applying Recommendation Technology. Maryam Ramezania,, Lawrence Bergmanb, Rich Thompsonb, Robin Burkea, Bamshad Mobashera. In Proceedings of International Workshop on Recommendation and Collaboration, in Conjunction with 2008 International ACM Conference on Intelligent User Interfaces (IUI 2008), Canaria, Canary Islands, Spain, January 2008. (PDF) "This paper presents a taxonomy of recommender systems with the goal of assisting in selection and application of these systems. Recommendation methods are usually classified into three main categories: collaborative, contentbased, and knowledge-based. We outline a taxonomy of recommender systems based on problem characteristics and the underlying technology. We show how the taxonomy can help researchers and developers select between different kinds of recommender systems by clearly defining the problem characteristics including: problem structure, domain, relationship with the user, user input, background knowledge, and recommendation outputs."

Tutorial Slides & Notes

AI Techniques for Personalized Recommendation John Riedl, Anthony Jameson, and Joseph Konstan. Full-day tutorial presented at AAAI 2004, the Nineteenth National Conference on Artificial Intelligence. San Jose, California, U.S.A., 26 July 2004. The slides for this tutorial are available in a single PDF file, with 2 slides per page.