流恋Chapter7
Context-Aware Recommender Systems
Gediminas Adomavicius and Alexander Tuzhilin
Abstract The importance of contextual information has been recognized by re-archers and practitioners in many disciplines,including e-commerce personal-ization,information retrieval,ubiquitous and mobile computing,data mining,mar-keting,and management.While a substantial amount of rearch has already been performed in the area of recommender systems,most existing approaches focus on recommending the most relevant items to urs without taking into account any ad-ditional contextual information,such as time,location,or the company of other ,for watching movies or dining out).In this chapter we argue that relevant contextual information does matter in recommender systems and that it is important to take this information into account when providing recommendations.We discuss the general notion of context and how it can be modeled in recommender systems. Furthermore,we introduce three different algorithmic paradigms–contextual pre-filtering,post-filtering,and modeling–for incorporating contextual information into the recommendation process,discuss the possibilities of combining veral context-aware recommendatio
n techniques into a single unifying approach,and provide a ca study of one such combined approach.Finally,we prent additional capabil-ities for context-aware recommenders and discuss important and promising direc-tions for future rearch.
Gediminas Adomavicius
Department of Information and Decision Sciences
Carlson School of Management,University of Minnesota
e-mail:gedas@umn.edu
Alexander Tuzhilin
Department of Information,Operations and Management Sciences
Stern School of Business,New York University
e-mail:u.edu
F. Ricci et al. (eds.), Recommender Systems Handbook,
217 DOI 10.1007/978-0-387-85820-3_7, © Springer Science+Business Media, LLC 2011
218Gediminas Adomavicius and Alexander Tuzhilin 7.1Introduction and Motivation
The majority of existing approaches to recommender systems focus on recommend-ing the most relevant items to individual urs and do not take into consideration any contextual information,such as time,place and the company of other , for watching movies or dining out).In other words,traditionally recommender sys-tems deal with applications having only two types of entities,urs and items,and do not put them into a context when providing recommendations.
However,in many applications,such as recommending a vacation package,per-sonalized content on a Web site,or a movie,it may not be sufficient to consider only urs and items–it is also important to incorporate the contextual information into the recommendation process in order to recommend items to urs under certain cir-cumstances.For example,using the temporal context,a travel recommender system would provide a vacation recommendation in the winter that can be very different from the one in the summer.Similarly,in the ca of personalized content delivery on a Web site,it is important to determine what content needs to be delivered(rec-ommended)to a customer and when.More specifically,on weekdays a ur might prefer to read world news when she logs on in conspiracy theory
the morning and the stock market report in the evening,and on weekends to read movie reviews and do shopping.
The obrvations are consistent with thefindings in behavioral rearch on consumer decision making in marketing that have established that decision making, rather than being invariant,is contingent on the context of decision making.There-fore,accurate prediction of consumer preferences undoubtedly depends upon the degree to which the recommender system has incorporated the relevant contextual information into a recommendation method.
More recently,companies started incorporating some contextual information into their recommendation engines.For example,when lecting a song for the customer, Sourcetone interactive radio()takes into the consideration the current mood of the listener(the context)that she specified.In ca of music recom-menders,some of the contextual information,such as listener’s mood,may matter for providing better recommendations.However,it is still not clear if context matters for a broad range of other recommendation applications.
In this chapter we discuss the topic of context-aware recommender systems (CARS),address this and veral other related questions,and demonstrate that,de-pending on the application domain and
the available data,at least certain contextual information can be uful for providing better recommendations.We also propo three major approaches in which the contextual information can be incorporated into recommender systems,individually examine the three approaches,and also discuss how the three parate methods can be combined into one unified ap-proach.Finally,the inclusion of the contextual information into the recommenda-tion process prents opportunities for richer and more diver interactions between the end-urs and recommender systems.Therefore,in this chapter we also discuss novelflexible interaction capabilities in the form of the recommendation query lan-guage for context-aware recommender systems.
vigina7Context-Aware Recommender Systems219 The rest of the chapter is organized as follows.Section7.2discuss the general notion of context as well as how it can be modeled in recommender systems.Sec-tion7.3prents three different algorithmic paradigms for incorporating contextual information into the recommendation process.Section7.4discuss the possibili-ties of combining veral context-aware recommendation techniques and provides a ca study of one such combined approach.Additional important capabilities for context-aware recommender systems are described in Section7.5,and the conclu-sions and some opportunities for future work are prented in Section7.6.
7.2Context in Recommender Systems
Before discussing the role and opportunities of contextual information in recom-mender systems,in Section7.2.1we start by discussing the general notion of con-text.Then,in Section7.2.2,we focus on recommender systems and explain how context is specified and modeled there.
7.2.1What is Context?
Context is a multifaceted concept that has been studied across different rearch dis-ciplines,including computer science(primarily in artificial intelligence and ubiqui-tous computing),cognitive science,linguistics,philosophy,psychology,and orga-nizational sciences.In fact,an entire conference–CONTEXT(e,for example, context-07.ruc.dk)–is dedicated exclusively to studying this topic and incor-porating it into various other branches of science,including medicine,law,and busi-ness.In reference to the latter,a well-known business rearcher and practitioner C. K.Prahalad has stated that“the ability to reach out and touch customers anywhere at anytime means that companies must deliver not just competitive products but also unique,real-time customer experiences shaped by customer context”and that this would be the next main issue(“big thing”)for the CRM practitioners[57].
Since context has been studied in multiple disciplines,each discipline tends to take its own idiosyncratic view that is somewhat different from other disciplines and is more specific than the standard generic dictionary definition of context as “conditions or circumstances which affect some thing”[70].Therefore,there ex-ist many definitions of context across various disciplines and even within specific subfields of the disciplines.Bazire and Br´e zillon[17]prent and examine150 different definitions of context from differentfields.This is not surprising,given the complexity and the multifaceted nature of the concept.As Bazire and Br´e zillon[17] obrve:
“...it is difficult tofind a relevant definition satisfying in any discipline.Is context a frame for a given object?Is it the t of elements that have any influence on the object?Is it possible to define context a priori or just state the effects a posteriori?Is it something static
220Gediminas Adomavicius and Alexander Tuzhilin or dynamic?Some approaches emerge now in Artificial Intelligence[...].In Psychology,we generally study a person doing a task in a given situation.Which context is relevant for our study?The context of the person?The context of the task?The context of the interaction?
The context of the situation?When does a context begin and where does it stop?What are the real relationships between context and cognition?”
enjoy的用法
Since we focus on recommender systems in this paper and since the general concept of context is very broad,we try to focus on thofields that are directly related to recommender systems,such as data mining,e-commerce personalization, databas,information retrieval,ubiquitous and mobile context-aware systems,mar-keting,and management.We follow Palmisano et al.[54]in this ction when de-scribing the areas.
Data Mining.In the data mining community,context is sometimes defined as tho events which characterize the life stages of a customer and that can determine a change in his/her preferences,status,and value for a company[18].Examples of context include a new job,the birth of a child,marriage,divorce,and retirement. Knowledge of this contextual information helps(a)mining patterns pertaining to this particular context by focusing only on the relevant data;for example,the data pertaining to the daughter’s wedding,or(b)lecting only relevant ,tho data mining results that are applicable to the particular context,such as the discov-ered patterns that are related to the retirement of a person.
E-commerce Personalization.Palmisano et al.[54]u the intent of a purcha made by a customer in an e-commerce application as contextual information.Dif-ferent purchasing intents may lead to different types of behavior.For example,the same customer may buy from the same online account d
ifferent products for dif-ferent reasons:a book for improving her personal work skills,a book as a gift,or an electronic device for her hobby.To deal with different purchasing intentions, Palmisano et al.[54]build a parate profile of a customer for each purchasing con-text,and the parate profiles are ud for building parate models predicting customer’s behavior in specific contexts and for specific gments of customers. Such contextual gmentation of customers is uful,becau it results in better predictive models across different e-commerce applications[54].
Recommender systems are also related to e-commerce personalization,since per-sonalized recommendations of various products and rvices are provided to the customers.The importance of including and using the contextual information in rec-ommendation systems has been demonstrated in[3],where the authors prented a multidimensional approach that can provide recommendations bad on contextual information in addition to the typical information on urs and items ud in many recommendation applications.It was also demonstrated by Adomavicius et al.[3] that the contextual information does matter in recommender systems:it helps to increa the quality of recommendations in certain ttings.
Similarly,Oku et al.[53]incorporate additional contextual dimensions(such as time,companion,and weather)into the recommendation process and u machine learning techniques to provide recomm
endations in a restaurant recommender sys-tem.They empirically show that the context-aware approach significantly outper-
7Context-Aware Recommender Systems221 forms the corresponding non-contextual approach in terms of recommendation ac-curacy and ur’s satisfaction with recommendations.
Since we focus on the u of context in recommender systems in this paper,we will describe the and similar approaches later in the chapter.
Ubiquitous and mobile context-aware systems.In the literature pertaining to the context-aware systems,context was initially defined as the location of the ur,the identity of people near the ur,the objects around,and the changes in the ele-ments[63].Other factors have been added to this definition subquently.For in-stance,Brown et al.[23]include the date,the ason,and the temperature.Ryan et al.[61]add the physical and conceptual status of interest for a ur.Dey et al.
梨的英文
怀疑英文ssd是什么[33]include the ur’s emotional status and broaden the definition to any informa-tion which can characterize and is relevant to the interaction between a ur and an application.Some associate the context with the ur[33,35],while others empha-size how context relates to the application[60,69].More recently,a number of other techniques for context-aware systems have been
discusd in rearch literature,in-cluding hybrid techniques for mobile applications[59,71]and graphical models for visual recommendation[20].
This contextual information is crucial for providing a broad range of Location-Bad Services(LBSes)to the mobile customers[64].For example,a Broadway the-ater may want to recommend heavily discounted theater tickets to the Time Square visitors in New York thirty minutes before the show starts(since the tickets will be wasted anyway after the show begins)and nd this information to the visitors’smart phones or other communication devices.Note that time,location,and the type of the communication ,smart phone)constitute contextual information in this application.Brown et al.[22]introduce another interesting application that allows tourists interactively share their sighteing experiences with remote urs, demonstrating the value that context-aware techniques can provide in supporting social activities.define是什么意思
A survey of context-aware mobile computing rearch can be found in[30], which discuss different models of contextual information,context-nsing tech-nologies,different possible architectures,and a number of context-aware application examples.
Databas.Contextual capabilities have been added to some of the databa man-agement systems
英语国家线
by incorporating ur preferences and returning different answers to databa queries depending on the context in which the queries have been ex-presd and the particular ur preferences corresponding to specific contexts.More specifically,in Stephanidis et al.[66]a t of contextual parameters is introduced and preferences are defined for each combination of regular relational attributes and the contextual parameters.Then Stephanidis et al.[66]prent a context-aware extension of SQL to accommodate for such preferences and contextual informa-tion.Agrawal et al.[7]prent another method for incorporating context and ur preferences into query languages and develop methods of reconciling and ranking different preferences in order to expeditiously provide ranked answers to contextual queries.Mokbel and Levandoski[52]describe the context-aware and location-aware
222Gediminas Adomavicius and Alexander Tuzhilin databa rver CoreDB and discuss veral issues related to its implementation, including challenges related to context-aware query operators,continuous queries, multi-objective query processing,and query optimization.
Information Retrieval.Contextual information has been proven to be helpful in information retrieval and access[40],although most existing systems ba their re-trieval decisions solely on queries and document collections,whereas information about arch context is often ignored[9].The effectiveness
of a proactive retrieval system depends on the ability to perform context-bad retrieval,generating queries which return context-relevant results[46,65].In Web arching,context is consid-ered as the t of topics potentially related to the arch term.For instance,Lawrence [45]describes how contextual information can be ud and propos veral special-ized domain-specific context-bad arch engines.Integration of context into the Web rvices composition is suggested by Maamar et al.[51].Most of the current context-aware information access and retrieval techniques focus on the short-term problems and immediate ur interests and requests(such as“find allfiles created during a spring meeting on a sunny day outside an Italian restaurant in New York”), and are not designed to model long-term ur tastes and preferences.
Marketing and Management.Marketing rearchers have maintained that the purchasing process is contingent upon the context in which the transaction takes place,since the same customer can adopt different decision strategies and prefer different products or brands depending on the context[19,50].According to Lilien et al.[47],“consumers vary in their decision-making rules becau of the usage sit-uation,the u of the good or rvice(for family,for gift,for lf)and purcha situation(catalog sale,in-store shelf lection,and sales person aided purcha).”Therefore,accurate predictions of consumer preferences should depend on the de-gree to which we have incorporated t
he relevant contextual information.In the mar-keting literature,context has been also studied in thefield of behavioral decision theory.In Lussier and Olshavsky[50],context is defined as a task complexity in the brand choice strategy.
微笑服务演讲稿
The context is defined in Prahalad[57]as“the preci physical location of a cus-tomer at any given time,the exact minute he or she needs the rvice,and the kind of technological mobile device over which that experience will be received.”Fur-ther,Prahalad[57]focus on the applications where the contextual information is ud for delivering“unique,real-time customer experiences”bad on this contex-tual information,as oppod to the delivery of competitive products.Prahalad[57] provides an example about the ca when he left his laptop in a hotel in Boston, and was willing to pay significant premiums for the hotel shipping the laptop to him in New York in that particular context(he was in New York and needed the laptop really urgently in that particular situation).
To generalize his statements,Prahalad[57]really distinguishes among the fol-lowing three dimensions of the contextual information:temporal(when to deliver customer experiences),spatial(where to deliver),and technological(how to de-liver).Although Prahalad focus on the real-time experiences(implying that it is
