Medium and long term runoff trend prediction method based on deep learning model

A medium and long term runoff trend prediction method based on a deep learning model comprises the following steps: S1, constructing a runoff comprehensive index representing a monthly runoff change trend and an influence object thereof, and obtaining an early observation value; S2, performing factor screening on the primary selection factors based on a partial mutual information method; S3, determining the network depth of the deep belief network model through a particle swarm algorithm, dynamically optimizing parameters such as the number of neurons of each hidden layer in the deep belief network model, the learning rate of an RBM and the learning rate finely adjusted by a BP algorithm by adopting a flower pollination algorithm, and constructing a PSO-FPA-DBN model; and S4, predicting a medium and long term runoff change trend based on the constructed PSO-FPA-DBN model so that comprehensive representation of the runoff change trend of the whole drainage basin is achieved. Apartial mutual information method