Load prediction method and system based on empirical mode decomposition and deep neural network

The invention discloses a load prediction method and system based on empirical mode decomposition and a deep neural network. The method comprises the following steps: acquiring initial power load data of a time sequence; performing empirical mode decomposition on the initial electrical load data to obtain a series of intrinsic mode functions and residual errors; converting the intrinsic mode function and the residual error into a data matrix; inputting the data matrix into a convolutional neural network model to obtain a feature vector; and inputting the feature vector and initial electrical load data of a time sequence into the long short-term memory network model to obtain predicted electrical load data. The initial electrical load data is decomposed into the intrinsic mode functions with different feature scales through empirical mode decomposition, so that the periodicity of the load data is more obvious, the feature extraction capability of the convolutional neural network and the long-short term memory