Analisis Kepatuhan Belanja Dana BOS Melalui Klasterisasi K-Means++: Studi Kasus Pada Satuan Pendidikan Tingkat Menengah Pertama
DOI:
https://doi.org/10.30865/jurikom.v13i1.9424Keywords:
K-Means++, Spending Compliance, BOS Funds, SIPLah, Clustering, TransparencyAbstract
The Ministry of Education, Culture, Research, and Technology Regulation Number 18 of 2022 mandates all educational units to procure goods and services using School Operational Assistance (BOS) funds through the SIPLah platform. This policy is a strategic government step to enhance transparency, accountability, and efficiency in the management of national education funds. However, field implementation still faces significant challenges characterized by low platform adoption rates, where the majority of schools tend to prioritize conventional manual procurement prone to administrative irregularity risks. This study aims to analyze the compliance level of BOS fund spending through spending pattern clustering to identify SIPLah adoption characteristics and formulate targeted policy recommendations. The method used is the K-Means++ algorithm, selected for its superiority in optimal centroid initialization compared to standard K-Means, resulting in more accurate clusters with fast convergence. Research data sourced from the 2024 School Activity Plan and Budget (RKAS) of 40 State Junior High School (SMPN) samples selected using Slovin's formula with a 10% error rate. The clustering process produced three distinct compliance characteristic groups: Cluster 1 (Low Compliance) dominating with 30 schools (75%) in the SIPLah usage range of 0-5%, Cluster 2 (Medium Compliance) comprising 8 schools (20%) in the range of >5-15%, and Cluster 3 (High Compliance) comprising only 2 schools (5%) in the range of >15%. Validation using the Silhouette Coefficient yielded a score of 0.687, indicating a well-formed cluster structure. These findings provide a significant contribution to the education office as a basis for policy evaluation, where schools in Cluster 1 require infrastructure audits and intensive assistance, while Cluster 3 can serve as a best practice model in school financial governance.
References
[1] M. Umdatul and S. A. Zaenab, "Attractive: Innovative Education Journal. Students' Difficulties at Elementary School in Increasing
Literacy Ability," Attractive Innov. Educ. J., vol. 4, no. 1, pp. 1–12, 2024.
[2] G. Sania, I. Zufria, and M. Fakhriza, "Penerapan Metode AHP dan SAW Pada Sistem Pendukung Keputusan Untuk Menentukan
Pengalokasian Dana BOS," J. Inf. Technol., vol. 4, no. 1, pp. 117–125, 2024, doi: 10.46229/jifotech.v4i1.869.
[3] K. M. Tahir, "Kebijakan Alokasi Pembagian Dana BOS Sesuai Jumlah Standar Minimal BOS Funding Allocation Policy According
to Minimum," vol. 13, pp. 79–98, 2023.
[4] D. S. Diantari, S. O. M. Bani, A. S. Fatiqh, and S. Trihantoyo, "Pengoperasian Aplikasi RKAS Dalam Pengelolaan Dana Bantuan
Operasional Sekolah (BOS)," J. Ilmu Pendidik., vol. 1, no. 2, pp. 280–299, 2024, doi: 10.62383/hardik.v1i2.370.
[5] D. Herinanto, M. Gumanti, and B. H. S. Utami, "Pemanfaatan SIPLah Dalam Memasarkan Produk Ekonomi Kreatif Bagi Para
Pengusaha Mikro, Kecil, dan Menengah di Desa Podosari," SWARNA J. Pengabdi. Kpd. Masy., vol. 3, no. 3, pp. 280–289, 2024, doi:
10.55681/swarna.v3i3.1247.
[6] Nugroho and F. D. Adhinata, "Penggunaan Metode K-Means dan K-Means++ Sebagai Clustering Data Covid-19 di Pulau
Jawa," Teknika, vol. 11, no. 3, pp. 170–179, 2022, doi: 10.34148/teknika.v11i3.502.
[7] Y. Li, "User Role Discovery and Optimization Method Based on K-means++ and Reinforcement Learning in Mobile
Applications," CMES - Comput. Model. Eng. Sci., vol. 131, no. 3, pp. 1365–1386, 2022, doi: 10.32604/cmes.2022.019656.
[8] C. V. Addad, S. Lattanzi, N.-A. Fard, C. Sohler, and O. Svensson, "Fast and accurate k-means++ via rejection sampling," Adv.
Neural Inf. Process. Syst., vol. 2020-Decem, no. NeurIPS, 2020.
