LSTM-Driven Forecasting of Surface Temperature Trends in Indonesia as Insights Into Climate Change

One of Indonesia's Sustainable Development Goals (SDGs) is to combat climate change. In the context of climate change, one of the most obvious signs is the increase in surface temperatures. The average world temperature has increased significantly since the 19th century, and it is crucial to comprehend this increase to comprehend how the plan to reach the 13th SGD can help manage future climate change. To compare the three approaches—Long Short-Term Memory (LSTM) Vanilla, Bi-directional LSTM (Bi-LSTM), and 2-Stacked LSTM— for surface temperature forecasting in Indonesia. With the smallest evaluation metric value from RMSE of 34.385, MAE of 27.84%, and MAPE of 1.11001%, the 2-Stacked LSTM model is the best, according to the results. The forecasting results show that in the next 2 years there is no significant change in the improvement of weather changes.