Load Forecasting Based on SABO-PSO-ELM Hybrid Algorithm

The short-term forecast of power load is of great significance to the planning and development of the power industry. In this chapter, we propose a new method to enhance the performance of predictive models by combining subtraction mean optimization (SABO), particle swarm optimization algorithm (PSO), and extreme learning machine (ELM). Through careful parameter adjustment and optimization process, this study successfully demonstrated the effectiveness of SABO and PSO algorithms in optimizing ELM parameters, which greatly improved the application performance of the model in practical prediction tasks. In addition, the method in this study provides a feasible solution for processing complex data sets, enhancing the adaptability and stability of the model in the face of data changes. Future work will explore the potential of this approach for other types of machine learning tasks and on larger data sets. The proposed method was evaluated using the 2021 load data of the PJM public dataset and the 2022 full-year data of an industrial park in Liaoning province. The SABO-PSO-ELM method is compared with other mainstream methods (BWO-ELM). The statistical analysis shows that the proposed method has better prediction accuracy on the four standard scales of MSE, MAPE, MAD, and NRMSE, which reflects the advanced nature of the method.