<![CDATA[Gold is a strategic investment instrument that serves as a hedge against inflation, although its price movements are characterized by high volatility and complex patterns. This study develops and compares the performance of the conventional statistical approach, Autoregressive Integrated Moving Average (ARIMA), and the deep learning approach, Recurrent Neural Network (RNN) and Neural Basis Expansion Analysis for Interpretable Time Series (N-BEATS), in forecasting the weekly average gold prices in Jakarta during the 2022–2024 period. The two methods are evaluated using the Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Symmetric Mean Absolute Percentage Error (sMAPE) metrics. The results indicate that N-BEATS outperformed ARIMA in forecasting performance. This is attributed to its basis expansion architecture, which effectively captures nonlinear dynamics and adapts to extreme volatility, making it more accurate and robust than RNN and ARIMA.]]>
Copyrights © 2026
