Research on Temperature Prediction Based on RF-LSTM Modeling

In this paper, a new method for temperature prediction combining Random Forest (RF) and Long Short-Term Memory Network (LSTM) is proposed to address the challenges in the field of weather prediction. The method combines the advantages of RF and LSTM, which considers the temporal correlation of meteorological data and captures the intrinsic correlation among meteorological elements. While LSTM mines the temporal patterns in the historical weather data, RF provides more informative input data, which improves the timeliness and accuracy of the prediction. Compared with the LSTM model, the RF-LSTM model has improved the RMSE, MAE, and MAPE values by <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$0.27,0.22$</tex>, and 0.07, respectively, and the RF-LSTM model performs significantly in the temperature prediction in Nanjing with high prediction accuracy and stability. The method provides a new idea for meteorological prediction, which not only improves the overall level of prediction, but also provides valuable guidance for people's production and life.