Customer-Side Load Prediction Based on ISSA-Optimized EEMD

To reduce the error in load forecasting results, this paper proposes an electricity load forecasting method based on a Hybrid Improved Sparrow Search Algorithm (ISSA) optimized Ensemble Empirical Mode Decomposition (EEMD). Historical user electricity consumption data is denoised using the Singular Value Decomposition (SVD) algorithm and smoothed via the Autoregressive Integrated Moving Average (ARIMA) model. An optimization objective function is established for the parameters required by EEMD, and the optimal parameter combination is derived based on the ISSA algorithm. By analyzing historical electricity load data, the processed data is fed into a forecasting model based on a Generalized Regression Neural Network (GRNN) to output load prediction values. Experimental results demonstrate that the relative error of the proposed method’s predictions remains within 0.1, meeting the requirements for user-side electricity load forecasting.