<![CDATA[Water is a fundamental human need that must be available to support daily activities. However, ensuring the availability of clean water remains a significant challenge for both communities and water providers, whether managed by the private sector or the government. One effective approach to ensuring the availability of clean water is accurately predicting the water discharge from raw water sources. The results of these predictions can be used to map the potential for water availability and identify areas with water deficits. This study focuses on predicting water discharge in the Kolong ST 12 reservoir, located in the Sungailiat District of Bangka Regency. The Singular Spectrum Analysis (SSA) method is used for prediction. The aim of this study is to evaluate the accuracy of the SSA method as a prediction model. The results show that predictions made using the SSA method are quite accurate, as evidenced by the Mean Absolute Percentage Error (MAPE) values of 46.68% for in-sample data and 35.54% for out-of-sample data. These MAPE values indicate that the SSA method has an accuracy of 53.32% for in-sample predictions and 64.46% for out-of-sample predictions. Parameter selection and data preprocessing processes affect the accuracy of the SSA method prediction results. The results of this study indicate that the SSA method has the potential to be developed into an accurate prediction model in analyzing water discharge data for optimal water resources management. The implications of this study are that the findings can assist the government and relevant stakeholders in water management. This study provides strategic information for policy planning to support food security programs and prevent water shortages through well-planned distribution.]]>
Copyrights © 2025
