<![CDATA[As a tropical country, Indonesia faces great challenges in predicting rainfall due to increasingly dynamic climate change. This study aims to predict rainfall in an urban area in West Java with tropical climate characteristics using deep learning methods, namely Bidirectional Long Short-Term Memory (Bi-LSTM) and Gated Recurrent Unit (GRU) based on climate data collected from local meteorological stations. The results show that the Bi-LSTM method provides more stable prediction performance with a Mean Absolute Error (MAE) value of 0.0108 and a Root Mean Squared Error (RMSE) of 0.0158. In contrast, the GRU method produced variable performance with higher MAE and RMSE values in some test scenarios. The main findings of this study indicate that the BiLSTM model has a higher level of accuracy, making it an effective information technology solution to support disaster mitigation and agricultural sector planning in climatically complex regions.]]>
Copyrights © 2025
