Context-aware recommender systems
Proceedings of the 2008 ACM conference on Recommender systems - RecSys '08
T raditional recommender systems, such as those based on content-based and collaborative filtering, tend to use fairly simple user models. For example, user-based collaborative filtering generally models the user as a vector of item ratings. As additional observations are made about users' preferences, the user models are extended, and the full collection of user preferences is used to generate recommendations or make predictions. This approach, therefore, ignores the notion of "situated
... of "situated actions" (Suchman 1987), the fact that users interact with the system within a particular "context" and that preferences for items within one context may be different from those in another context. In many application domains, a context-independent representation may lose predictive power because potentially useful information from multiple contexts is aggregated. For example, when a user is buying books, the preferences the user expresses in one context, such as "books for my children," may be of no predictive value when the user seeks recommendations in a different context, such as "work-related books." The ideal contextaware recommendation system would, therefore, be able reliably to label each user action with an appropriate context and effectively tailor the system output to the user in that given context. The concept of "context" has been studied extensively in several areas of computing and other disciplines. For example, Bazire and Brezillon (2005) examine and compare some 150 different definitions of context from a number of different fields and conclude that the multifaceted nature of the concept makes it difficult to find a unifying definition: "Is context a frame for a given object? Is it the set of elements that have any influence on the object? Is it possible to define context a priori or just state Articles FALL 2011 67 n Context-aware recommender systems (CARS) generate more relevant recommendations by adapting them to the specific contextual situation of the user. This article explores how contextual information can be used to create intelligent and useful recommender systems. It provides an overview of the multifaceted notion of context, discusses several approaches for incorporating contextual information in the recommendation process, and illustrates the usage of such approaches in several application areas where different types of contexts are exploited. The article concludes by discussing the challenges and future research directions for context-aware recommender systems.