<![CDATA[Stability in the economy is influenced by technological advancements, which impact the digitization of the economy and lead to an increasing demand for electronic and digital payment systems compared to physical currency. There are certain months, such as during year-end holidays, when the circulation of physical currency increases. This study purpose to forecasting the total currency circulation in Indonesia, considering the influence of calendar variations, using a hybrid method that combines SARIMAX and NN. The SARIMAX method was utilized to capture linear effects related to calendar variations, while the NN method was employed to capture nonlinear patterns. The analysis results indicated that the hybrid SARIMAX-NN model with 1 to 3 neurons yielded accurate forecasts, with Mean Absolute Percentage Error (MAPE) values below 2%. However, the highest accuracy was achieved by the SARIMAX-NN hybrid model with 1 neuron, which had a MAPE of 1.38%. Additionally, the forecasting results showed a consistent monthly increase, particularly during the holiday season in December]]>
Copyrights © 2023
