Machine Learning and Spatial Analysis of Territorial Determinants For Coverage Optimization In Online Higher Education

This study proposes a comprehensive analytical framework for the strategic positioning of Hybrid Support Centers (HSCs) aimed at mitigating geographic, technological, and socioeconomic barriers in online higher education. Drawing on socioeconomic data from students represented through mixed-type variables, the framework integrates weighted Gower distance with unsupervised learning techniques (agglomerative hierarchical clustering and PAM refinement) and supervised models for large-scale profile prediction, reaching a classification accuracy of 78.67%. Results are spatially projected through territorial indicators to identify priority intervention zones. The principal contribution lies in a reproducible clustering and predictive framework for mixed data, integrated with territorial planning tools, that promotes more equitable and inclusive access to higher education.