A Study on Monthly Precipitation Prediction in Chengdu City Based on QPSO-LSTM

In this paper, a precipitation prediction model based on QPSO-LSTM is proposed to optimize the hyperparameter combinations of the LSTM network through the QPSO algorithm to improve the prediction accuracy and efficiency of the model. The QPSO algorithm enhances the global searching ability, accelerates the speed of convergence, and avoids the local optimal solution efficiently, which makes the model have a significant advantage in capturing the seasonal changes of precipitation. Based on the historical precipitation data of Chengdu city, the experiment proves that the QPSO-LSTM model is superior to the traditional LSTM and PSO-LSTM models in terms of accuracy, which is of great theoretical and practical significance for future precipitation prediction.