Improving the Academic Environment and Teaching Quality through Data Segmentation and Decision Trees

Abstrak

This study aims to assess and segment graduate student satisfaction in the Master of Management Program at UMB, focusing on academic services and teaching quality, using a data-driven approach through K-Means clustering and Decision Tree classification. The K-Means method successfully identified three distinct satisfaction clusters: Very Satisfied, Satisfied, and Dissatisfied. From an HRD perspective, this segmentation provides strategic insights into how students perceive institutional service delivery—spanning administrative processes, supporting infrastructure, and interaction quality across service units. These insights reflect the internal customer experience, which is critical in managing human capital performance within higher education institutions. The clustering results were further examined using a Decision Tree to identify key attributes driving satisfaction levels. Factors such as clarity in professional certification procedures (K55), responsiveness of certification services (K56), and management service quality (K78) emerged as dominant predictors of satisfaction in the service domain. In the teaching domain, lecturer empathy (K03) and responsiveness (K01) were shown to significantly influence student engagement and academic trust—two core components in HRD models for faculty performance. These findings suggest the need for targeted capacity-building programs, soft skill enhancement, and infrastructure development. By integrating student feedback into ongoing HR practices and quality assurance mechanisms, institutions can foster a more human-centered, responsive, and effective educational environment.