<![CDATA[This study compares the performance of Long Short-Term Memory (LSTM) and Random Forest (RF) models in predicting temperature data from Tanjung Priok, Indonesia, using evaluation metrics such as RMSE, MAE, and R² Score. The LSTM model demonstrated its ability to capture temporal dependencies and temperature trends, achieving an R² score of 0.4493 and an MAE of 0.5863. In contrast, the RF model performed better in minimizing prediction errors, with a lower RMSE of 0.6498 and an R² score of 0.4066. While the LSTM model excelled in explaining variance in the temperature data, the RF model was more effective in stable periods, exhibiting lower prediction errors. The results highlight that both models have distinct advantages, with LSTM better suited for capturing long-term temperature trends and RF performing well during periods of stability. Future research could explore hybrid models or further optimization of these techniques to improve prediction accuracy, particularly for dynamic and extreme temperature fluctuations.]]>
Copyrights © 2025
