A new Frontier in electric load forecasting: The LSV/MOPA model optimized by modified orca predation algorithm

• DOI: 10.1016/j.heliyon.2024.e24183
• Publication Date: 2024-01-12T04:29:54Z
• Authors: Guanyu Yan; Jinyu Wang; Myo Thwin
• Keywords: Electrical load

Electric load forecasting is a vital task for energy management and policy-making. However, it also challenging problem due to the complex dynamic nature of electric data. In this paper, novel technique, called LSV/MOPA, has been proposed forecasting. The technique hybrid model that combines advantages Long Short-Term Memory (LSTM) Support Vector Regression (SVR), two powerful artificial intelligence algorithms. further optimized by newly Modified Orca Predation Algorithm (MOPA), which enhances accuracy efficiency. LSV/MOPA applied historical data from South Korea, covering four regions 20 years. compared with other state-of-the-art techniques, including SVR/FFA, LSTM/BO, LSTM-SVR, CNN-LSTM. results show minimum average mean absolute percentage deviation error, 365 in northern region, 12.8 southern 8.6 central 30.8 eastern provides best fitting outperforms techniques terms Mean Absolute Percentage Deviation (MAPD) index, achieving lower values all exhibits faster convergence better generalization than techniques. This study demonstrates effectiveness superiority suggests its potential applications sectors where accurate crucial.