POWER LOAD FORECASTING METHOD BASED ON LONG SHORT-TERM MEMORY NEURAL NETWORK

摘要

A power load forecasting method based on a long short-term memory neural (LSTM) network. The method comprises the steps of: inputting power load data and a region feature factor at a historical moment by means of an input unit of a computer (S21); training and modeling the power load data and the region feature factor at the historical moment by means of an LSTM network, in order to generate a deep neural network load forecasting model by training (S22), the deep neural network load forecasting model being a single-layer multi-task deep neural network model or a double-layer multi-task deep neural network model used for power supply load forecasting; forecasting the power load in a region needing to be forecasted by means of the deep neural network load forecasting model generated by training, and generating a forecasting result of the power load in the region (S23); and outputting the forecasting result of the power load in the region by means of an output unit of the computer (S24). By constructing a power load forecasting model for multi-task learning on the basis of an LSTM network in the deep learning field, power consumption loads in multiple regions can be precisely forecasted and the forecasting effect is improved.