Electric power load prediction method and prediction device based on LSTM network model

The invention discloses an electric power load prediction method and prediction device based on an LSTM network model, and the method comprises the following steps: obtaining a data set for prediction, including historical electric power load power data, meteorological data and economic data, and screening a sample data set; preprocessing the sample data set, and performing data standardization and data missing value processing on the time sequence data and the power load data to obtain each unit time load feature data set; the constructed LSTM network model comprises an input layer, a hidden layer and an output layer; and training the LSTM network model, and optimizing model parameters by using a gradient descent method to minimize a loss function. And performing prediction by using the trained LSTM network model, and inputting future meteorological data, economic data and the like. According to the method, various influence factors can be abstracted, a long-time sequence can be processed, the long-term depe