Power grid electric quantity prediction method based on data mining and neural network

The invention relates to a power grid electric quantity prediction method based on data mining and a neural network, and the method comprises the following steps: firstly, data cleaning and recovery: employing a 3 sigma principle to remove abnormal points, and employing secondary Lagrange interpolation to recover missing data, and avoiding the influence of the missing data on time continuity; then, constructing electric quantity characteristics from internal and external factors, considering historical electric quantity data for the internal factors, combining the historical electric quantity data with time coding, considering factors such as regional population and economy for the external factors, and introducing the factors into a prediction model; and finally, establishing a prediction model based on a Prophet model and an LSTM model, carrying out component decomposition on the power grid electric quantity data, respectively establishing LSTM networks for different components, carrying out fitting on the