SXRS: An XLink-based Recommender System using Semantic Web technologies

Traditional recommender systems adopt a static view of the recommendation process and treat it as a prediction problem. We argue that it is more appropriate to view the problem of generating recommendation lists as a dynamic link generation and it is feasible to apply the XML Link Language (XLink) for a novel recommendation solution. However, XLink lacks knowledge and semantic representation to cope with the computer-interpretable effects. In this paper, we address the issue developing an Semantic XLink Recommendation System (SXRS), which is based on semantic web technologies and is composed of XLink base, knowledge base, Search Engine, and Inference Engine to provide three different approaches to represent the linking knowledge – XLink-based metadata, ontology-based reasoning, and rule-based inference. To illustrate the SXRS applications, we implemented an academic domain recommender system that offers dynamic recommendation lists to assist student to plan semester courses.