Chapter

Recommender Systems Handbook

pp 611-648

Human Decision Making and Recommender Systems

  • Anthony JamesonAffiliated withDFKI, German Research Center for Artificial Intelligence Email author 
  • , Martijn C. WillemsenAffiliated withEindhoven University of Technology
  • , Alexander FelfernigAffiliated withUniversity of Graz
  • , Marco de GemmisAffiliated withDepartment of Computer Science, University of Bari “Aldo Moro”
  • , Pasquale LopsAffiliated withDepartment of Computer Science, University of Bari “Aldo Moro”
  • , Giovanni SemeraroAffiliated withDepartment of Computer Science, University of Bari “Aldo Moro”
  • , Li ChenAffiliated withHong Kong Baptist University

* Final gross prices may vary according to local VAT.

Get Access

Abstract

If we assume that an important function of recommender systems is to help people make better choices, it follows that people who design and study recommender systems ought to have a good understanding of how people make choices and how human choice can be supported. This chapter starts with a compact synthesis of research on the various ways in which people make choices in everyday life, in terms of six choice patterns; we explain for each pattern how recommender systems can support its application, both in familiar ways and in ways that have not been explored so far. Similarly, we distinguish six high-level strategies for supporting choice, noting that one strategy is directly supported by recommendation technology but that the others can also be applied fruitfully in recommender systems. We then illustrate how this conceptual framework can be used to shed new light on several fundamental questions that arise in recommender systems research: In what ways can explanations of recommendations support choice processes? What are we referring to when we speak of a person’s “preferences”? What goes on in people’s heads when they rate an item? What is “choice overload”, and how can recommender systems help prevent it? How can recommender systems help choosers to engage in trial and error? What subtle influences on choice can arise when people choose among a small number of options; and how can a recommender system take them into account? One general contribution of the chapter is to generate new ideas about how recommendation technology can be deployed in support of human choice, often in conjunction with other strategies and technologies.