Elsevier

Expert Systems with Applications

Volume 42, Issue 3, 15 February 2015, Pages 1743-1758
Expert Systems with Applications

Review
A systematic review of scholar context-aware recommender systems

https://doi.org/10.1016/j.eswa.2014.09.017Get rights and content

Highlights

  • We review the relevant articles in the field of scholar recommendations.

  • We explore contextual information influential in scholar recommendations.

  • We examine recommending approaches.

  • Contextual information are categorised in three groups.

  • The most recommending approaches are collaborative filtering, content based, knowledge based and hybrid.

Abstract

Incorporating contextual information in recommender systems is an effective approach to create more accurate and relevant recommendations. This review has been conducted to identify the contextual information and methods used for making recommendations in digital libraries as well as the way researchers understood and used relevant contextual information from the years 2001 to 2013 based on the Kitchenham systematic review methodology. The results indicated that contextual information incorporated into recommendations can be categorised into three contexts, namely users’ context, document’s context, and environment context. In addition, the classical approaches such as collaborative filtering were employed more than the other approaches. Researchers have understood and exploited relevant contextual information through four ways, including citation of past studies, citation of past definitions, self-definitions, and field-query researches; however, citation of the past studies has been the most popular method. This review highlights the need for more investigations on the concept of context from user viewpoint in scholarly domains. It also discusses the way a context-aware recommender system can be effectively designed and implemented in digital libraries. Additionally, a few recommendations for future investigations on scholarly recommender systems are proposed.

Introduction

Recommender Systems (RSs) have been an area of substantial research interest since the mid-1990s (Felfernig & Burke, 2008). In the last decade, RSs had been investigated and implemented in various application domains, including knowledge management, e-commerce, e-learning and e-health (Verbert, Lindstaedt, & Gillet, 2010).

The dramatic data increase in Digital Libraries (DLs) has necessitated the use of RSs as an appropriate tool for facilitating and accelerating the process of information seeking (Porcel & Herrera-Viedma, 2010). Scientists prefer to have most of their required information at their fingertips. They usually input keywords to retrieve the desired scientific information in DLs, but the results may not always be what they would expect. Hence, the retrieval of relevant information has been a time-consuming task for most of them. Consequently, providing proper information is a significant factor for an effective DL in a scientific environment. Libraries try to apply intelligent personalised systems such as RSs (Mönnich & Spiering, 2008) to support users by offering relevant resources based on their interests and preferences (Sikka, Dhankhar, & Rana, 2012). RSs can manage information overload by filtering and personalising data according to users’ needs (Adomavicius et al., 2005, Pommeranz et al., 2012); thus, RSs normally collect data about users’ activities and build user models to filter the preferences expressed either explicitly or implicitly (Baltrunas, Ludwig, Peer, & Ricci, 2012).

In recent years, RSs use the information describing users’ situations such as location, time, and task in order to generate more relevant and personalised recommendations (Adomavicius and Tuzhilin, 2011, Asabere, 2013). For example, the resources recommended to an undergraduate student searching for “Fuzzy method” for his class assignment may be different from those recommended to a graduate student writing a research paper on the same topic. This is due to the different requirements of the tasks they are working on and the different levels of formal education, which are considered as contextual information.

Using contextual information has been considered as a main source of accuracy of recommendations (Adomavicius and Tuzhilin, 2011, Baltrunas, 2008). Researchers emphasise applying contextual approaches in order to recommend items to users based on certain circumstances (Baltrunas and Ricci, 2009, Kaminskas and Ricci, 2011). However, the variety of application scenarios and user requirements cause difficulties in presenting an unanimous definition of contextual information for all Context-Aware Recommender Systems (CARS) (Yujie & Licai, 2010). Moreover, to predict accurate recommendations for users of a specific domain such as DLs, it is essential to understand and exploit the relevant contexts of users, which lead to creating intelligent recommendations. Therefore, the aim of this study is to carry out a literature review on RSs for the academic DLs in order to:

  • (a)

    Identify the contextual information that has been adopted for making recommendations in the academic DLs.

  • (b)

    Identify the approaches that have been used to adopt contextual information for making recommendations in the academic DLs.

  • (c)

    Explore how the relevance of contextual information to recommendations for an academic domain has been understood by researchers before applying it.

We conducted this review based on the guidelines by Kitchenham and Charters (2007) for performing systematic literature reviews in software engineering. We explain more about the methodology of our review in Section 4. The rest of the paper is organised as follows. We discuss a few definitions of context from various points of views and provide recommendation approaches in Section 2. The related works are presented in Section 3. The methodology of this study is presented in Section 4. We report and discuss the results from performing the review in Section 5. The results are structured according to the research questions.

Section snippets

What is context?

Many definitions of context have been proposed in various disciplines, including computer science (primarily in artificial intelligence and ubiquitous computing,), information retrieval, cognitive science, linguistics, philosophy, social science, psychology, and organisational sciences (Adomavicius & Tuzhilin, 2011); it is beyond the scope of this research to review all of them. However, from a general point of view, the Oxford Advanced Learner’s Dictionary mentions that context is “a situation

Related work

Although there are numerous studies on context-aware systems, there is no systematic review on CARSs in academic DLs as well as no study to review contextual information incorporated into recommendations in academic DLs. Nonetheless, a few reviews in the field of RSs in DLs or classifications of RSs in DLs can be helpful in identifying the need for a systematic review in this area. Hence, we discuss them as below.

A literature review conducted by (Park, Kim, Choi, & Kim, 2012) examined research

Methodology

We used the systematic literature review methodology introduced by Kitchenham and Charters (2007) since it is a rigorous and well-defined method in the fields of software engineering. As the word systematic indicates, systematic literature review aims to specify questions and review relevant studies in order to identify gaps in the current research, as well as appraise their contributions to questions and gaps for drawing conclusion in a particular research question, area, or phenomenon. As

Results

In this section, we report the results (Phase 3: reporting the review) of this systematic review and discuss them in order to respond to the three questions posed in Section 1.

Discussion and recommendations

Scholar recommender systems are employed to facilitate the process of information seeking for users. Incorporating contextual information in recommender systems is an effective approach to create more accurate and relevant recommendations. In this paper, we first discuss the concept of context and contextual approaches; we then examined 82 papers published on scholarly recommender systems from the years 2001 to 2013 by using the Kitchenham systematic review methodology in order to identify

Acknowledgements

This research is supported by UM High Impact Research Grant UM.C/625/1/HIR/MOE/FCSIT/05 from the Ministry of Higher Education Malaysia and the University of Malaya under University of Malaya Research Grant RP003B-13ICT.

References (152)

  • G. Adomavicius et al.

    Toward the next generation of recommender systems: A survey of the state-of-the-art and possible extensions

    IEEE Transactions on Knowledge and Data Engineering

    (2005)
  • G. Adomavicius et al.

    Incorporating contextual information in recommender systems using a multidimensional approach

    ACM Transactions on Information Systems (TOIS)

    (2005)
  • S. Aftab et al.

    An approach for secure semantic data integration at data as a service (DaaS) layer

    International Journal of Information and Education Technology

    (2015)
  • M. Aittola et al.

    SmartLibrary–location-aware mobile library service

    Human-computer interaction with mobile devices and services

    (2003)
  • B. Amini et al.

    Incorporating scholar’s background knowledge into recommender system for digital libraries

  • N.Y. Asabere

    Towards a viewpoint of context-aware recommender systems (CARS) and services

    International Journal of Computer Science and Telecommunications

    (2013)
  • M. Balabanović et al.

    Fab: Content-based, collaborative recommendation

    Communications of the ACM

    (1997)
  • L. Baltrunas et al.

    Context-based splitting of item ratings in collaborative filtering

  • L. Baltrunas

    Exploiting contextual information in recommender systems

  • L. Baltrunas et al.

    Context relevance assessment and exploitation in mobile recommender systems

    Personal and Ubiquitous Computing

    (2012)
  • C. Basu et al.

    Technical paper recommendation: A study in combining multiple information sources

    Journal of Artificial Intelligence Research

    (2011)
  • M. Bazire et al.

    Understanding context before using it

  • Beel, J., Docear, M., Langer, S., Genzmehr, M., Gipp, B., Breitinger, C., et al. (2013). Research paper recommender...
  • J. Beel et al.

    A comparative analysis of offline and online evaluations and discussion of research paper recommender system evaluation

  • C. Bizer et al.

    Linked data-the story so far

    International Journal on Semantic Web and Information Systems (IJSWIS)

    (2009)
  • U. Bojārs et al.

    Linked data for digital libraries

  • J. Bollen et al.

    An architecture for the aggregation and analysis of scholarly usage data

  • P.J. Brown et al.

    Context-aware applications: From the laboratory to the marketplace

    IEEE Personal Communications

    (1997)
  • R. Burke

    Hybrid recommender systems: Survey and experiments

    User Modeling and User-Adapted Interaction

    (2002)
  • R. Burke

    Hybrid web recommender systems

  • P.G. Campos et al.

    Context-aware movie recommendations: An empirical comparison of pre-filtering, post-filtering and contextual modeling approaches

  • K. Chandrasekaran et al.

    Concept-based document recommendations for citeseer authors

  • C.-M. Chen et al.

    An intelligent mobile location-aware book recommendation system with map-based guidance that enhances problem-based learning in libraries

  • R.-S. Chen et al.

    Using data mining to provide recommendation service

    WSEAS Transactions on Information Science and Applications

    (2008)
  • Q. Cheng Li et al.

    Research of information recommendation system based on reading behaviour

  • L. Costabello et al.

    Context-aware access control for RDF graph stores

  • E. Craig
    (1998)
  • J. Crowther

    Oxford advanced learner’s dictionary

    (1995)
  • M.R. De Giusti et al.

    An ontology-based context aware system for selective dissemination of information in a digital library

    Journal of Computing

    (2010)
  • Z. Dehghani et al.

    A multi-layer contextual model for recommender systems in digital libraries

    Aslib Proceedings

    (2011)
  • M. Deshpande et al.

    Item-based top-n recommendation algorithms

    ACM Transactions on Information Systems (TOIS)

    (2004)
  • A.K. Dey

    Understanding and using context

    Personal and Ubiquitous Computing

    (2001)
  • S. Dietze et al.

    A survey on linked data and the social web as facilitators for TEL recommender systems

  • P. Dourish

    What we talk about when we talk about context

    Personal and Ubiquitous Computing

    (2004)
  • A. Felfernig et al.

    Constraint-based recommender systems: technologies and research issues

  • M. Franke et al.

    Recommender services in scientific digital libraries

  • Z. Gantner et al.

    MyMediaLite: A free recommender system library

  • G. Geisler et al.

    Developing recommendation services for a digital library with uncertain and changing data

  • A. Geyer-Schulz et al.

    An architecture for behavior-based library recommender systems

    Information Technology and Libraries

    (2003)
  • A. Geyer-Schulz et al.

    Others also use: A robust recommender system for scientific libraries

  • Cited by (110)

    • How do context-aware artificial intelligence algorithms used in fitness recommender systems? A literature review and research agenda

      2022, International Journal of Information Management Data Insights
      Citation Excerpt :

      While emerging technologies are pertinent for improving recommendation output efficiency, they also impose challenges that must be resolved. Context-aware recommender systems have been viewed from various perspectives (Villegas et al., 2018; Verbert et al., 2012; Champiri et al., 2015). But prior work has insisted on the need to identify the contextual information and challenges in various application domains.

    • New contextual collaborative filtering system with application to personalized healthy nutrition education

      2022, Journal of King Saud University - Computer and Information Sciences
    View all citing articles on Scopus
    View full text