Reservoir level prediction via integrating long short-term memory and graph structure learning
The water level change of reservoirs is affected by many factors such as rainfall,flood discharge,and evaporation.The prediction accuracy of existing reservoir water level prediction methods needs to be im-proved.Therefore,a reservoir water level prediction model was proposed integrating long short-term memory(LSTM)and graph convolution neural network(GCN).The proposed model first extracts time-series depend-ent features of water level and related influencing factors by using LSTM.Then,a graph structure learning module is designed to automatically capture the correlation between water level and different influencing factors.Finally,GCN is used for feature learning and prediction.Extensive experiments were conducted on the Three Gorges Dam dataset and datasets provided by cooperative enterprises.The experimental results demon-strated the effectiveness and superiority of the proposed model.
NETL
NSTL
万方数据
