Joint prediction method and device for regional cooling, heating and power loads

The invention relates to a joint prediction method and device for regional cooling, heating and power loads. The method comprises the steps: obtaining the historical data of the regional cooling, heating and power loads, and carrying out noise reduction preprocessing; performing correlation analysis on the data, determining influence factors, and dividing the data into a training set, a test set and a verification set; obtaining a pre-constructed neural network set, which comprises a basic LSTM neural network structure, a stacked LSTM network structure and a bidirectional LSTM network structure, and adopting the training set to train each network structure until a preset network convergence condition is reached; then verifying errors of each network structure by using the verification set, and finally performing verification prediction by using the verification set; and selecting the network structure with the optimal prediction result as the optimal network structure, and predicting the regional cooling, hea