Comprehensive energy electricity-cold-heat load prediction method and device and storage medium

The invention relates to the technical field of energy load prediction, in particular to a comprehensive energy electricity-cold-heat load prediction method and device and a storage medium, and the method comprises the steps: collecting meteorological data and load data, and carrying out the preprocessing of the data; decomposing and reconstructing the preprocessed data into three groups of time sequence data with a high frequency band, a middle frequency band and a low frequency band by utilizing a wavelet decomposition algorithm; based on a causal correlation analysis method, related meteorological factor data and power-cold-heat load data of the integrated energy system are screened out; the method comprises the following steps: constructing an electricity-cold-heat load prediction model based on a partial decoupling extended Kalman filtering neural network, training the neural network by adopting a partial decoupling extended Kalman filtering algorithm, and training the load prediction model; and using th