Cluster algorithm method for profile analysis of scientific researchers
Guardado en:
Autor: | |
---|---|
Formato: | artículo original |
Estado: | Versión publicada |
Fecha de Publicación: | 2022 |
Descripción: | The increase in scientific production makes it a challenge to identify particular patterns and traits that characterize researchers. Establishing levels of compatibility and similarity between actors in a scientific research context from their profiles requires a rapid and appropriate process. The objective of this article is to evaluate the levels of similarity, Euclidean distance and compatibility between vectors of researchers, based on clustering algorithms, multidimensional scaling, principles of the vector-space model and attributes of their scientific profiles, considering the terminologies addressed in their scientific production. Theoretical and empirical methods were used, including text mining techniques and tools. The application of the procedure in the Advanced Energy and Technology Study Center from Cuba and the Cotopaxi Technical University from Ecuador, evidenced its effectiveness. As a result, it was possible to identify professionals with higher levels of coincidence in areas and lines of research, which favors the establishment of Collective Communities of Knowledge; it was possible to demonstrate that the methods used can be integrated to ICT, resulting in obtaining perceptual relationships between researchers and expressing the groups formed from clusters of observations in each subcategory and knowledge domains of the two case studies analyzed. |
País: | Portal de Revistas UCR |
Institución: | Universidad de Costa Rica |
Repositorio: | Portal de Revistas UCR |
Lenguaje: | Español |
OAI Identifier: | oai:portal.ucr.ac.cr:article/50456 |
Acceso en línea: | https://revistas.ucr.ac.cr/index.php/eciencias/article/view/50456 |
Palabra clave: | cluster analysis user profiles vector space model análisis de conglomerados perfiles de usuario modelo de espacio vectorial |