%0 Thesis
%9 Skripsi
%A Wati, Ildia
%A Universitas Pakuan,
%A Fakultas Matematika dan Ilmu Pnegetahuan Alam,
%A Program Studi Ilmu Komputer,
%B Fakultas Matematika dan Pengetahuan Alam
%D 2026
%F eprintsunpak:11228
%I Universitas Pakuan
%T Decision Support System for Prioritizing Extreme Poverty   Alleviation at the Provincial Level in Indonesia Using K-Means   and Random Forest Algorithms
%U http://eprints.unpak.ac.id/11228/
%X Decision Support System for Prioritizing Extreme Poverty   Alleviation at the Provincial Level in Indonesia Using K-Means   and Random Forest Algorithms  Ildia Wati  Computer Science Study Program, Faculty of Mathematics and Natural Sciences,  Pakuan University, Bogor City, West Java, 16143, Indonesia  Abstract  Extreme poverty remains one of the major challenges in Indonesia’s national development   agenda, with disparities in the severity of poverty across provinces making it difficult to   objectively determine government intervention priorities. This study aims to develop a decision   support system for identifying priority provinces for extreme poverty alleviation in Indonesia   using a hybrid approach that combines K-Means Clustering and Random Forest. Panel data   obtained from Statistics Indonesia (BPS) for the period 2016–2025 were utilized, comprising   34 provinces from 2016 to 2023 and 38 provinces from 2024 to 2025, resulting in 348   observations. Seven indicators were employed: poverty rate, Gini ratio, Human Development   Index (HDI), Open Unemployment Rate (OUR), Economic Growth Rate (EGR), Poverty Gap   Index (P1), and access to improved sanitation. K-Means Clustering was first applied to   generate three objective priority labels (low, medium, and high), which were subsequently used   as training labels for the Random Forest classification model. Model performance was   evaluated using 10-fold Group Cross-Validation to prevent data leakage in the panel data   structure, achieving an accuracy of 92.82%. Feature importance analysis revealed that the   Poverty Gap Index (P1) was the most influential indicator (0.238), followed by the poverty rate   and the Gini ratio, while the Economic Growth Rate contributed the least. The resulting model   was implemented in a Laravel-based web application, with a pure PHP inference service to   ensure that the classification results remained consistent with the model trained in Python. The   proposed system provides an objective, data-driven decision support tool to assist   policymakers and researchers in prioritizing extreme poverty alleviation programs across   Indonesian provinces.  Keywords: extreme poverty; K-Means clustering; Random Forest; decision support system; provincial   prioritizatio