Analysis of co-authorship networks among Brazilian graduate programs in computer science

TL;DR

This study analyzes co-authorship networks among Brazilian computer science graduate programs to identify topological patterns and relate them to program quality evaluations by CAPES.

Contribution

It presents a novel analysis of co-authorship network structures of Brazilian graduate programs and correlates them with qualitative evaluation scores.

Findings

Identified key topological measures characterizing programs.

Revealed patterns correlating network features with CAPES evaluations.

Provided insights into research collaboration structures in Brazilian academia.

Abstract

The growth and popularization of platforms on scientific production have been the subject of several studies, producing relevant analyses of coauthorship behavior among groups of researchers. Researchers and their scientific productions can be analyzed as coauthorship social networks, so researchers are linked through common publications. In this context, coauthoring networks can be analyzed to find patterns that can describe or characterize them. This work presents the analysis and characterization of co-authorship networks of academic Brazilian graduate programs in computer science. To this end, data from the curricula of Brazilian researchers were collected and modeled as coauthoring networks among the graduate programs that researchers participate in. Each network topology was analyzed regarding complex network measurements and three qualitative indices that evaluate the publications quality. In addition, the coauthorship networks of the graduate programs were characterized in relation to the evaluation received by CAPES, which attributes a qualitative grade to the graduate programs in Brazil. The results indicate some of the most relevant topological measures for the programs characterization and evaluate at different qualitative rates and indicate a pattern of the graduate programs best evaluated by CAPES.