An Architecture for Scholarly Recommender System Based on Identified Contextual Information in Medical Sciences

Today, researchers prefer to have most of their required information at their fingertips. Scholarly or research paper recommender systems are intelligent systems that aim to recommend the most appropriate scientific papers or resources based on user's needs. Past studies have shown that contextual information such as users', system' and environment' contexts influence the quality of recommendations. Therefore, the goal of this research is to identify effective user-oriented contextual information which influences the process of recommendation to scholars in medical area and then to present an architecture to design and develop an scholarly recommender system.
Semi-structured interviews were carried out with 50 medical science professors and PhD studies in order to identify contextual information. Data resulted from interviews were analyzed in three stages using open coding, followed by axial and selective coding, developed in the Grounded Theory methodology. Then, contextual information has been exploited for a multi-layer architecture in design of a scholarly recommender system.
The results of our data analysis showed that scholars’ attributes such as purposes, literacy, skills, mental status, suppositions and assumptions, occupational condition, and social status are among the most influential factors which should be considered in designing a scholarly recommender system. Finally, based on the findings, we designed a multilayer system.
Exploitation of contextual information in intelligent systems such as recommender systems and search engines leads to a better interaction between users and systems; therefore, the results of this study can be beneficial for designing other systems in this area.