Fig 1.
A visual architecture of LSTM-RNN.
Fig 2.
Training of decision tree.
Fig 3.
Testing of input data v in a decision tree.
Fig 4.
Flowchart depicting the development process of the proposed stacking ensemble model in the current research.
Fig 5.
Rainfall distribution in Bangladesh, 1963–2022.
Fig 6.
Humidity distribution in Bangladesh, 1963–2022.
Fig 7.
Average temperature distribution in Bangladesh, 2022.
Fig 8.
Sea level pressure in Bangladesh, 2022.
Fig 9.
Histogram and Q-Q Plots of A) Rainfall B) Humidity C) Sea Level Pressure D) Average Temperature.
Fig 10.
Normal probability plot of temperature.
Fig 11.
Analysis of meteorological data using heat map.
Fig 12.
Training and validation loss of RNN-LSTM model.
Table 1.
Results for RNN-LSTM layer architecture.
Fig 13.
Average temperature forecast by the RNN-LSTM model versus the actual value in 2022.
Fig 14.
Graph illustrating prediction error against the number of trees.
Fig 15.
Construction of a TensorFlow decision forest plot model.
Fig 16.
Average temperature forecast by the TFDF model versus the actual value in 2022.
Table 2.
Evaluation criteria for TFDF model.
Table 3.
Criteria for evaluating three basic models.
Table 4.
Evaluating the stacking prediction model.
Fig 17.
Average temperature forecast by the Stacking model versus the actual value in 2022.
Fig 18.
Future average temperature prediction for Bangladesh (2025).