Water level data early warning method for key water level stations in Yangtze River basin

The invention discloses a water level data early warning method for key water level stations in the Yangtze River basin, which comprises the following steps: completing related configuration linked with a database, and integrating the water level data of the water level stations; environment variables and related configurations suitable for software and an algorithm are adjusted, water level data variables of adjacent sites are defined, algorithm functions to be realized are programmed on vscode by using a python language, mutual information values corresponding to different grid intervals are calculated by using a maximum information coefficient (MIC) and a Pearson return coefficient of a big data set algorithm, and the mutual information values of the different grid intervals are calculated; analyzing the correlation of adjacent water level stations in two scenes (flood season and non-flood season); applying a neural network algorithm to carry out noise addition and training on the data, and expressing a re