Recommender Systems with Social Elements: A Systematic Mapping

Recommendation Systems (RS) deal of the overload of information online, allowing the user to find desirable items quickly, without being surprised by irrelevant information. Individual preference and interpersonal influence are important contextual factors for social recommendations, as they affect users' decisions about information retention. The goal of this paper is to identify the state of the art in RS with the use of social elements. For this, a systematic mapping of the literature was conducted, revealing a growing trend in the number of articles published in the last ten years, especially in China, and with more frequent proposals for new models, systems and frameworks for recommendation. Almost half of the mapped articles present as a domain Entertainment or Product Review/Evaluation, with the collaborative filtering approach being the most common of the approaches used, and the similarity of friends as the most common of the social components considered. As an evaluation strategy, more than half of the mapped articles use offline experiments in a previously populated database to simulate user actions. The mapping showed that although RSs are considering social elements, there is still a lack of works that explore these elements in real contexts of use.