Monthly precipitation prediction in Luoyang city based on EEMD-LSTM-ARIMA model

Abstract

Abstract At present, the method of using coupled models to model different frequency subseries of precipitation series separately for prediction is still lacking in the research of precipitation prediction, thus in this paper, a coupled model based on Ensemble Empirical Mode Decomposition (EEMD), Long Short-Term Memory neural network (LSTM) and Autoregressive Integrated Moving Average (ARIMA) is proposed for month-by-month precipitation prediction. The monthly historical precipitation data of Luoyang City from 1973 to 2021 were used to build the model, and the modal components of different frequencies obtained by EEMD decomposition were divided into high-frequency series part and low-frequency series part using the Permutation Entropy (PE) algorithm, the LSTM model is used to predict the high-frequency sequence part, while the ARIMA model is used to predict the low-frequency sequence part. Monthly precipitation forecasts are obtained by superimposing the results of the two models. Finally, the predictive performance is evaluated using several assessment metrics. The indicators show that the model predictive performance outperforms the EMD-LSTM (Empirical Mode Decomposition), EEMD-LSTM, EEMD-ARIMA combined models and the single models, and the model has high confidence in the prediction results of future precipitation.