Recommender Systems with Side Information

in Special Issue   Posted on November 3, 2017 

Information for the Special Issue

Special Issue Call for Papers:

Purpose. Recommender systems research has made significant advances over the past decades and has seen wide adoption in electronic commerce. Recently, a variety of types of side information (e.g., social friends, item content) has been incorporated into recommender systems to further enhance their performance, especially the well-recognized problem of data sparsity. However, most of existing approaches have only investigated the value of a single type of side information at a time, such as social trust, friendship, or item contents. In real-life applications, users may have different kinds of reactions towards items of interest. For example, users often search and compare several alternative products, click and view product details, zoom in and out product images, and so on, before they purchase one of these products. Another example is that users form different types of communities in their social networks, including those based on friendship ties, and others based on common interests or behaviors. Even the context of user-item interactions is often multi-dimensional though, including temporal, geographical, social and weather information. The multiple relations among users, items and related data impose new challenges for the researchers. It is necessary to build new theories, techniques and methods to exploit multi-dimensional (homogeneous and heterogeneous) side information to provide users with better personalized recommendations. At the same time, the large volume and variety of side data and the velocity of incremental updates in live systems provide challenges for the scalable mining and application of user preferences.

Topics. In this Special Issue, we invite papers focusing on the integration of multiple types of side information for recommender systems within e-commerce applications. They include those that take advantage of the correlations between side information and target data, exploit the characteristics of side information to model user preference and provide explanations to generated recommendations, and tackle the problems of scalability and variety of side information. Together, these support new theories, techniques, approaches and toolkits for next-level recommender systems.

The issue will focus on technologies and solutions related, but not limited to:

  • Algorithms that exploit homogenous and heterogeneous side information to make better top-N item recommendations, and that address data sparsity and cold start issues.
  • Systems that can apply side information to enhance recommendation diversity, novelty and explainability.
  • Context-aware recommendation systems incorporating side information.
  • Large-scale parallelization techniques to speed up information fusion and generation of personalized recommendations
  • Incremental recommendation solutions to handle continuous updates, especially real-time streaming data for recommendations.
  • Innovative, efficient recommendation tools and benchmarks to integrate multiple types of side information and enhance reproducibility and comparison of such recommendation models.

Submission Window. The submission window for full papers is from July 1, 2017 to June 30, 2018. Papers submitted to ECRA prior to the open window that are connected in some way to the themes of the Special Issue will also be considered for it. Authors are welcome to submit abstracts to the Special Issue Editors, as a means to gauge interest and get suggestions on the development of a Special Issue paper.

Submission Guidelines. Only unpublished research papers will be considered. Prior published and indexed conference papers cannot be considered due to the copyright and IP concerns of ECRA’s publisher, Elsevier, unless they are significantly extended in content. Authors should limit their initial submissions to no more than 32 double-spaced pages in 11-point font with appropriate margins. References, figures, tables and appendices can add to that length. Author names and affiliations should be listed on the first page of the paper; the reviewing will be single blind only. All papers should be submitted via ECRA’s submission system at (Please note that ECRA is scheduled to cut over to a new reviewing system sometime in 2017, so the submission URL may change.)

Authors should select the “Recommender System Special Issue” tab when they reach the “Article Type” step in the submission process. Authors may request a sample paper from the Special Issue Editor for a fully formatted sample. Authors should follow Elsevier’s Electronic Commerce Research and Applications manuscript format, which can be found on the journal’s homepage at

Reviewing Process. Submissions are not batched for reviewing, but instead will be handled under review as they are received. ECRA’s Associate Editor (AE) and Guest Editors (GE) will assist with the reviewing process. The Special Issue Ges will try to return first reviews no later than 90 days from the date of an author’s first submission. ECRA uses a developmental reviewing approach for Special Issues, with the aim of helping authors to achieve very high quality final publications. Second and third round reviews, as needed, will be completed on an expedited basis, if authors are able to turn their revisions around quickly. Special Issue authors will be given an indication as early as possible of rejection for the Special Issue, including on the basis of a first reading by the Special Issue Editors. Inappropriately targeted or under-developed papers will be returned immediately with feedback.

Publication. The publication of Special Issue papers occurs as soon as possible after their completion using a method called ‘Article-Based Publication’ (ABP). The collation of such papers into a ‘Virtual Special Issue’ (VSI) using hyperlinks brings the content together. This ensures the rapid publication of individual papers. Our experience suggests that acceptance occurs between 6 to 12 months after a paper’s initial submission, and if it is accepted, it will reach publication in just a month .

Special Issue Editorial Team. This special issue will be coordinated by an Editorial Board Senior Editor, with the help of an Associate Editor and two Guest Editors, plus an Editorial Assistant:

  • Jie Zhang (ECRA SE, Coordinator) (Nanyang Tech. Univ., Singapore,
  • Hui Fang (ECRA AE) (Shanghai Univ. of Fin. and Econ., China,
  • Robin Burke (ECRA Guest Editor) (DePaul University, USA,
  • Guibing Guo (ECRA Guest Editor) (Northeastern Univ., China,
  • Emmy Hoang (ECRA’s Editorial Assistant for Special Issues) (for review process questions)

Interested authors should feel free to direct questions to Jie Zhang who will serve as the main contact and coordinating editor, or Emmy Hoang, who will be familiar with process and submission issues.

Other Special Issues on this journal

Closed Special Issues