<![CDATA[This study aims to predict Indonesia's oil and gas (migas) export values using the Seasonal Autoregressive Integrated Moving Average (SARIMA) and Long Short-Term Memory (LSTM) methods. Time series data from Statistics Indonesia (BPS) was utilized to develop an optimal prediction model. The selected SARIMA model, SARIMA(1,1,1)(1,1,1,12), was chosen based on the lowest Akaike Information Criterion (AIC) value. Meanwhile, the LSTM model was developed to capture more complex patterns in time series data. The forecasting results indicate that the SARIMA model provides higher accuracy compared to LSTM based on the Mean Absolute Percentage Error (MAPE), although LSTM demonstrated lower Mean Squared Error (MSE) and Root Mean Squared Error (RMSE). This study emphasizes that the choice of forecasting model should align with the characteristics of the data, where SARIMA is more suitable for oil and gas export data with seasonal patterns. These forecasting results can be utilized to support economic policy planning, optimize investments in the oil and gas sector, and mitigate global market fluctuation risks.]]>
Copyrights © 2025
