Clustering Achievement Records and Socioeconomic Profiles of KIP Kuliah Recipients

Riyadi Purwanto (Politeknik Negeri Cilacap)
Fajar Mahardika (Politeknik Negeri Cilacap)
Dwi Novia Prasetyanti (Politeknik Negeri Cilacap)
Cahya Vikasari (Politeknik Negeri Cilacap)

Journal of Embedded Systems, Security and Intelligent Systems
Vol 7 No 3 (2026): September 2026

Purpose – This study aims to identify patterns in the achievement records and socioeconomic characteristics of KIP Kuliah recipients to support institutional monitoring, differentiated student development, and administrative data-quality evaluation. Design/methods/approach – A quantitative unsupervised learning approach was employed using two complementary analytical streams. Achievement records were processed using text preprocessing, Term Frequency–Inverse Document Frequency (TF-IDF), Truncated Singular Value Decomposition, and K-Means clustering, with cluster selection and stability assessment used to evaluate the resulting segmentation. Socioeconomic characteristics of verified recipients were analyzed using Density-Based Spatial Clustering of Applications with Noise (DBSCAN), supported by parameter-sensitivity and alternative-distance analyses. Findings – The achievement analysis revealed distinct segments reflecting district-level academic and sports achievements, general district-level championships, national-level and olympiad achievements, and diverse sports and non-academic accomplishments. However, the achievement segments showed limited separation and should therefore be interpreted as descriptive patterns rather than definitive natural groupings. Socioeconomic clustering revealed heterogeneous recipient profiles and a substantial group of observations that did not form dense patterns. Sensitivity analysis further demonstrated that apparent low-income clusters were primarily driven by incomplete administrative records rather than verified economic vulnerability. Research implications/limitations – The findings support descriptive monitoring, differentiated achievement development, and improved verification of administrative data. Individual-level integration was not possible because the achievement and socioeconomic datasets lacked a common identifier. Originality/value – This study demonstrates how complementary text and socioeconomic clustering can reveal both meaningful recipient patterns and critical data-quality limitations in scholarship monitoring.