Incorporating Contextual Information in
Recommender Systems Using a
Multidimensional Approach
GEDIMINAS ADOMAVICIUS
University of Minnesota
RAMESH SANKARANARAYANAN
University of Connecticut
SHAHANA SEN
Fairleigh Dickinson University
and
ALEXANDER TUZHILIN
New York University
The article presents a multidimensional (MD) approach to recommender systems that can provide
recommendations based on additional contextual information besides the typical information on
users and items used in most of the current recommender systems. This approach supports multiple
dimensions, profiling information, and hierarchical aggregation of recommendations. The article
also presents a multidimensional rating estimation method capable of selecting two-dimensional
segments of ratings pertinent to the recommendation context and applying standard collaborative
filtering or other traditional two-dimensional rating estimation techniques to these segments. A
comparison of the multidimensional and two-dimensional rating estimation approaches is made,
and the tradeoffs between the two are studied. Moreover, the article introduces a combined rating
estimation method, which identifies the situations where the MD approach outperforms the stan-
dard two-dimensional approach and uses the MD approach in those situations and the standard
two-dimensional approach elsewhere. Finally, the article presents a pilot empirical study of the
combined approach, using a multidimensional movie recommender system that was developed for
implementing this approach and testing its performance.
Authors’ addresses: G. Adomavicius, Department of Information and Decision Sciences, Carlson
School of Management, University of Minnesota, 321 19th Avenue South, Minneapolis, MN 55455;
email: gedas@umn.edu; R. Sankaranarayanan, Department of Operations and Information Man-
agement, School of Business, University of Connecticut, 2100 Hillside Road, U-1041 OPIM, Storrs,
CT 06269-1041; email: rsankaran@business.uconn.edu; S. Sen, Department of Marketing and En-
trepreneurship, Silberman College of Business, Fairleigh Dickinson University, 1000 River Road
HDH2-07, Teaneck, NJ 07666; email: sen@fdu.edu; A. Tuzhilin, Department of Information, Op-
erations and Management Sciences, Stern School of Business, New York University, 44 West 4th
Street, New York, NY 10012; email: atuzhili@stern.nyu.edu.
Permission to make digital or hard copies of part or all of this work for personal or classroom use is
granted without fee provided that copies are not made or distributed for profit or direct commercial
advantage and that copies show this notice on the first page or initial screen of a display along
with the full citation. Copyrights for components of this work owned by others than ACM must be
honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers,
to redistribute to lists, or to use any component of this work in other works requires prior specific
permission and/or a fee. Permissions may be requested from Publications Dept., ACM, Inc., 1515
Broadway, New York, NY 10036 USA, fax: +1 (212) 869-0481, or permissions@acm.org.
C©
2005 ACM 1046-8188/05/0100-0103 $5.00
ACM Transactions on Information Systems, Vol. 23, No. 1, January 2005, Pages 103–145.

104
•
G. Adomavicius et al.
Categories and Subject Descriptors: H.1.2 [Models and Principles]: User/Machine Systems—
Human information processing; H.3.3 [InformationStorage andRetrieval]: Information Search
and Retrieval—Information filtering, Selection process
General Terms: Design, Algorithms, Experimentation, Performance
Additional Key Words and Phrases: Recommender systems, collaborative filtering, personalization,
multidimensional recommender systems, context-aware recommender systems, rating estimation,
multidimensional data models
1. INTRODUCTION AND MOTIVATION
There has been much work done in the area of recommender systems over the
past decade since the introduction of the first papers on the subject [Resnick
et al. 1994; Hill et al. 1995; Shardanand and Maes 1995]. Most of this work
has focused on developing new methods of recommending items to users and
vice versa, such as recommending movies to Web site visitors or recommending
customers for books. These recommendation methods are usually classified into
collaborative, content-based, and hybrid methods [Balabanovic and Shoham
1997] and are described in more detail in Section 2.
However, in many applications, such as recommending vacation packages,
personalized content on a Web site, various products in an online store, or
movies, it may not be sufficient to consider only users and items—it is also im-
portant to incorporate the contextual information of the user’s decision scenario
into the recommendation process. For example, in the case of personalizing
content on a Web site, it is important to determine what content needs to be de-
livered (recommended) to a customer and when. More specifically, on weekdays
a user might prefer to read world news in the morning and the stock market
report in the evening, and on weekends she might prefer to read movie reviews
and do shopping. As another example of the need for incorporating contextual
information, a “smart” shopping cart providing real-time recommendations to
shoppers using wireless location-based technologies [Wade 2003] needs to take
into account not only information about products and customers but also such
contextual information as shopping date/time, store, who accompanies the pri-
mary shopper, products already placed into the shopping cart and its location
within the store. Again, a recommender system may recommend a different
movie to a user depending on whether she is going to see it with her boyfriend
on a Saturday night or with her parents on a weekday. In marketing, behavioral
research on consumer decision making has established that decision making,
rather than being invariant, is contingent upon the context of decision making;
the same consumer may use different decision-making strategies and prefer
different products or brands under different contexts [Lussier and Olshavsky
1979; Klein and Yadav 1989; Bettman et al. 1991]. Therefore, accurate predic-
tion of consumer preferences undoubtedly depends upon the degree to which
relevant contextual information is incorporated into a recommendation method.
To provide recommendations incorporating contextual information, we
present a multidimensional recommendation model (MD model) that makes
recommendations based on multiple dimensions and, therefore, extends the
ACM Transactions on Information Systems, Vol. 23, No. 1, January 2005.