<![CDATA[Poverty in Indonesia requires precise analysis based on socio-economic indicators. This study develops classification models using Naïve Bayes, K-Nearest Neighbor (KNN), and Support Vector Machine (SVM). The primary focus is addressing class imbalance through the SMOTE technique. Utilizing 2021 BPS data from 515 regencies, the research incorporates 13 indicators, including education and infrastructure. Models were evaluated using accuracy, precision, recall, and F1-score across multiple data-split scenarios. Results indicate that SMOTE significantly enhances Naïve Bayes and KNN performance in identifying minority classes by reducing data bias. Conversely, SVM maintained consistent performance across all scenarios without SMOTE, attributed to its robust margin-based separation mechanism against distribution shifts. Overall, integrating SMOTE with machine learning algorithms improves classification reliability. This provides a crucial data-driven foundation for the government to formulate more targeted and equitable poverty alleviation policies across Indonesia, ensuring resources are allocated to the regions that need them most.]]>
Copyrights © 2026
