Enhancing the analysis of academic genealogy networks using dimensionality reduction techniques: A research internship at TU/e

• Grant number: 25/01858-6
• Support Opportunities: Scholarships abroad - Research Internship - Scientific Initiation
• Start date: April 30, 2025
• End date: August 29, 2025
• Field of knowledge: Physical Sciences and Mathematics - Computer Science - Computing Methodologies and Techniques
• Principal Investigator: Jesús Pascual Mena Chalco
• Grantee: Gabriel de Oliveira Dalpian
• Supervisor: Fernando Vieira Paulovich
• Host Institution: Centro de Matemática, Computação e Cognição (CMCC). Universidade Federal do ABC (UFABC). Santo André , SP, Brazil
• Institution abroad: Eindhoven University of Technology (TU/e), Netherlands
• Associated to the scholarship: 24/08240-5 - Detection and Interpretation of Outliers Using the LookOut Method in Academic Genealogy Graphs, BP.IC

Abstract

In recent years, the study of academic genealogy has emerged as a critical tool for understanding the dissemination of knowledge, mentorship patterns, and intellectual connections among researchers across generations. However, analyzing these networks remains challenging due to their inherent complexity and high dimensionality. This research integrates dimensionality reduction techniques, such as t-SNE and UMAP, into the LookOut framework, an advanced method for detecting, explaining, and visualizing outliers in academic genealogy datasets. By transforming high-dimensional data into lower-dimensional spaces, these techniques reveal patterns and relationships that are often obscured in traditional analyses. Building on the dataset from the Acácia Platform, which contains information on over 1.4 million Brazilian researchers, this project introduces a novel pipeline that leverages parameters, such as Descendance, Fecundity, and Genealogical Index, to uncover subtle anomalies within academic networks. A comparative evaluation of t-SNE and UMAP will assess their differences and enhance the LookOut framework's capabilities. Conducted in collaboration with the Research Group at Eindhoven University of Technology (TU/e), led by Prof. Fernando Paulovich, and the Scientometrics Research Group at Universidade Federal do ABC (UFABC), led by Prof. Jesús P. Mena-Chalco, this project bridges scientometric analysis and visualization techniques while fostering long-term collaboration. The outcomes are expected to advance network science and data visualization, culminating in a peer-reviewed publication highlighting methodological innovations and new insights into academic networks. (AU)