Um framework orientado a artigos para análise semântica automática de pesquisas sobre COVID-19

In this work, we propose a framework that automatically extracts semantic topics from scientific publications related to research on COVID-19. The framework has four main building blocks: (i) preprocessing, (ii) topic modeling, (iii) topic correlation with authors and institutions, and (iv) summarization interface. The first block corresponds to the application of pre-processing strategies in texts on the considered articles and the definition of their semantic representation. The topic modeling block aims to fi nd semantic topics in the publications used. The third block correlates these topics with the articles themselves and the authors, institutions, and countries related to each article. The summary interface provides an intuitive view for all these analyses. Our evaluation shows that our framework is capable of automatically extracting relevant characteristics from the articles, identifying the main themes addressed by them, as well as the correlation of researchers, institutions and countries for diff erent topics of research on COVID-19.