Análise das redes brasileiras de coautoria nos programas de pós-graduação em ciências da computação por meio de medidas topológicas
The analysis of social networks has become an area of great attention and focus in recent years, as it can be observed the behavior between its components, as well as their interactions. Co-authoring networks are an example of a social network, in which a researcher has a connection with another res...
| Autor principal: | Silva, Alex Junior Nunes Da |
|---|---|
| Formato: | Dissertação |
| Idioma: | Português |
| Publicado em: |
Universidade Tecnológica Federal do Paraná
2020
|
| Assuntos: | |
| Acesso em linha: |
http://repositorio.utfpr.edu.br/jspui/handle/1/5433 |
| Tags: |
Adicionar Tag
Sem tags, seja o primeiro a adicionar uma tag!
|
| Resumo: |
The analysis of social networks has become an area of great attention and focus in recent years, as it can be observed the behavior between its components, as well as their interactions. Co-authoring networks are an example of a social network, in which a researcher has a connection with another researcher as they share co-authoring in published articles, on these networks, topological measures can be applied to investigate patterns, classify and predict their behavior. In this work, the analysis of the Brazilian postgraduate programs in Computer Science was carried out. To this end, academic curricula were extracted from the Lattes Platform, mapping the connections between researchers and generating the representation of connections through graphs. It was adopted several topological measurements in order to evaluate the graphs. A quantitative index “Average Researchers per Publications” to measure program productivity was proposed, three qualitative indices of academic collaboration were also proposed, called “First Author Index”, “Collaboration Index”, and “Seniority Index” which analyze the author position in a publication and gives it a rating. The analyzed measurements were compared with the government’s periodic evaluation of the programs (CAPES Note) to validate their effectiveness. For comparison, classification approaches such as Random Forest, feature selection such as Best First and correlation were adopted. The results indicate a high relevance and identify patterns of behavior among the programs that explains the government’s evalution of the graduate programs. |
|---|