Short-term load forecasting using a two-stage CPO-PSO hyperparameter optimization of LSSVM.

Area of Science:

• Electrical Engineering
• Computational Intelligence
• Data Science

Background:

• Accurate short-term electricity load forecasting is essential for stable power system operation.
• Grid complexity and the dynamic nature of load data present forecasting challenges.

Purpose of the Study:

• To develop a novel hybrid model for enhanced short-term electricity load forecasting.
• To improve predictive accuracy and generalization capability in load prediction.

Main Methods:

• A hybrid model integrating Crested Porcupine Optimization (CPO) for global hyperparameter optimization, Particle Swarm Optimization (PSO) for local refinement, and Least Squares Support Vector Machine (LSSVM) for nonlinear modeling.
• Comparative analysis against baseline LSSVM, PSO-LSSVM, and CPO-LSSVM models using real-world datasets (Jiangsu and Australian).

Main Results:

• The CPO-PSO-LSSVM model demonstrated significant improvements in forecasting accuracy on both datasets.
• Reductions in Mean Absolute Error (MAE) and Root Mean Square Error (RMSE) were observed, alongside substantial increases in the Nash-Sutcliffe Efficiency (NSE).
• Specifically, on the Jiangsu dataset, MAE and RMSE decreased by 40.2% and 50.3%, with NSE rising to 0.902. The Australian dataset showed MAE and RMSE reductions of 25.3% and 50.8%, with NSE increasing to 0.996.

Conclusions:

• The proposed CPO-PSO-LSSVM model offers superior forecasting accuracy and robustness compared to existing methods.
• The hybrid approach is highly effective for short-term electricity load forecasting across different regions and time granularities.
• The model presents practical applicability for modern power system management and grid operation.