<![CDATA[Clean water availability is a key indicator of sustainable development, particularly in developing countries like Indonesia. Factors such as population growth, climate change, and urbanization contribute to fluctuations in clean water supply. This study aims to estimate the potential for clean water production in Indonesia using various machine learning algorithms, such as Linear Regression, Decision Tree, Random Forest, Multilayer Perceptron, XGBoost (Extreme Gradient Boosting), and Neural Network. Each algorithm was evaluated based on Mean Squared Error (MSE), Mean Absolute Error (MAE), R-squared (R²), and prediction accuracy. The results show that Linear Regression achieved the lowest MSE (9.31E-18), nearly zero, indicating extremely accurate predictions. Neural Network and Multilayer Perceptron also performed well, with MSE values of 0.00010898 and 0.00018004, respectively. Moreover, Linear Regression and Neural Network achieved R² scores of 1 and 0.9905, suggesting they can explain nearly all variability in the target data. These findings highlight the effectiveness of Linear Regression, Neural Network, and Multilayer Perceptron in modeling clean water production capacity. Therefore, these algorithms are recommended as the most reliable approaches for supporting data-driven decisions in clean water resource planning and management in Indonesia.]]>
Copyrights © 2025
