Nova Base de Dados Brasileira para Sistemas de Recomendação de Artigos Científicos

Abstract

This work presents a new dataset for Scientific Article Recommenda- tion Systems (SARS). In addition to being scarce, many existing datasets in the SARS context rely solely on co-authorship relationships as a relevance criterion, overlooking the importance of explicit user evaluations. To address this issue, we propose a new dataset with explicitly defined relevance labels, comprising over 2,000 researchers, 30 areas of knowledge, and approximately 71,000 as- sociated papers. The work includes a characterization and evaluation of the proposed dataset, alongside other widely used datasets in the literature