Penerapan Algoritma Machine Learning untuk Prediksi Tingkat Kemiskinan Berdasarkan Data Sosial Ekonomi Regional di Indonesia

Abstract

Poverty remains a major socio-economic challenge in Indonesia’s development. Various government programs have been implemented to reduce poverty rates, yet policy effectiveness is often constrained by limited capacity to accurately identify vulnerable regions and populations. The advancement of information technology and the increasing availability of socio-economic data provide opportunities to apply data science approaches for more precise decision-making. This study aims to implement machine learning algorithms to predict poverty levels based on regional socio-economic data in Indonesia. The dataset includes indicators such as unemployment rate, average years of schooling, education participation rate, inflation, Gross Regional Domestic Product (GRDP), population density, access to healthcare services, and other welfare indicators obtained from official statistical sources. A quantitative approach is employed through stages of data collection, preprocessing, feature selection, model development, and performance evaluation. The algorithms tested include Random Forest, Decision Tree, Support Vector Machine (SVM), and Gradient Boosting. Model performance is evaluated using accuracy, precision, recall, and F1-score. The results show that Random Forest achieves the best performance with higher accuracy and strong generalization. Key influencing factors include education level, unemployment rate, per capita income, and healthcare access. The model provides reliable predictions to support evidence-based poverty reduction policies.