Enhancing Governance Maturity Assessment in Higher Education Institutions Through Data-Driven Feature Selection and Clustering Techniques

Abstract

Various governance work steps need to be implemented continuously and harmoniously in higher education institutions so that academic quality can be maintained, regulatory standards can be met, accountability can be improved, and efficiency and alignment of organizational processes can be realized such as BAN-PT, ISO 9001:2015, and COBIT 2019. Although each framework has different characteristics and values, their simultaneous implementation often results in overlapping indicators, creates operational inefficiencies, and results in inconsistent maturity assessments. Although each framework has distinct characteristics and values, their simultaneous application often results in overlapping indicators, operational inefficiencies, and inconsistent maturity assessments. To address these challenges, this study develops an integrated approach that integrates framework harmonization processes, data-driven feature selection, and clustering techniques to produce a more accurate, efficient, and easily interpretable maturity assessment. This integration is designed to improve accuracy, efficiency, and interpretability in the preparation of governance maturity profiles. By utilizing a dataset consisting of 28 harmonized indicators collected from 15 academic units, we benchmarked three feature selection methods: Correlation-Based Feature Selection (CFS), Principal Component Analysis (PCA), and Information Gain (IG). The clustering algorithms (K-Means and DBSCAN). Validation using internal metrics (Silhouette = 0.62, DBI = 0.41, CHI = 342.5) indicates that the combination of CFS and K-Means yields the most valid and interpretable clusters, while retaining the semantic integrity of the original organizational structure unlike PCA, which produces abstract components. This approach facilitates the creation of actionable maturity profiles, benchmarking, and decision support in higher education administration.