<![CDATA[Institutional growth in higher education relies heavily on understanding applicant demographics and behavioral patterns to optimize recruitment strategies. This study presents a hybrid data mining pipeline combining Self-Organizing Maps (SOM) and K-Means clustering to segment student applicant data. SOM is utilized to project high-dimensional demographic and admission attributes onto a lower-dimensional topological space, while K-Means partitions the mapped structures into distinct, actionable segments. The clustering quality is rigorously evaluated using the Davies–Bouldin Index (DBI), where a minimum non-negative DBI value indicates optimal inter-cluster separation and intra-cluster cohesion. Empirical results on applicant records reveal distinct target profiles based on geographic origin, prior educational background, chosen study programs, and information channels. These findings provide university management with descriptive intelligence to tailor targeted marketing campaigns and resource allocation, replacing non-targeted recruitment practices with data-driven strategic planning.]]>
Copyrights © 2026
