Multi-station joint forecast model of water level in Hongze lake based on PSO-LightGBM

A multi-Station joint forecast model of water Level based on PSO-LightGBM is proposed to solve the problems of low prediction accuracy and short effective prediction period of traditional time series model and univariate time series model. The most relevant hydrological stations are found out by calculating the Pearson correlation coefficient between the water level of Hongze Lake and the water level of relevant hydrological stations. The PSO-LightGBM single variable water level prediction model of these relevant hydrological stations is established to predict the future water level of relevant hydrological stations. Together with the water level of Hongze Lake, it is used as the multivariable in the Hongze Lake water level prediction model. The multi-Station joint forecast model of water Level in Hongze Lake based on PSO-LightGBM is established to predict the water level sequence of the next ten days of Hongze Lake. The case study shows that: the model has higher prediction accuracy than univariate PSO-LightGBM, ARIMA and LSTM models.