<![CDATA[Palm oil production in Riau Province fluctuates due to seasonal factors and complex temporal patterns, making accurate forecasting with conventional statistical methods challenging. This study aims to compare the performance of six deep learning architectures Long Short-Term Memory (LSTM), Bidirectional LSTM (Bi-LSTM), Stacked LSTM, Gated Recurrent Unit (GRU), Bidirectional GRU (Bi-GRU), and Stacked GRU in predicting monthly palm oil production and identifying the most effective model. Two datasets obtained from PT Perkebunan Nusantara (PTPN) and Badan Pusat Statistik (BPS), covering the period from 2014 to 2023, were employed in this study. All models were evaluated using Mean Squared Error (MSE), Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE). The experimental results demonstrate that the GRU model consistently outperformed the other architectures, achieving the highest prediction accuracy of 95.84% and the lowest MAPE of 4.16% on the BPS dataset, while also producing the best overall performance on the PTPN dataset. These findings indicate that the research objective was successfully achieved and demonstrate that GRU provides a reliable, computationally efficient approach for forecasting palm oil production with limited time-series data. ]]>
