A quadratic $$\nu $$-support vector regression approach for load forecasting

• DOI: 10.1007/s40747-024-01730-7
• Publication Date: 2025-01-04T09:30:26Z
• Authors: Yanhe Jia; Shuaiguang Zhou; Yiwen Wang; Fengming Lin; Zheming Gao
• Keywords: Kernel-free support vector regression; Electronic computers. Computer science; Machine learning; Feature weighting; QA75.5-76.95; Information technology; Electric load forecasting; T58.5-58.64; Weighted support vector regression

This article focuses on electric load forecasting, which is a challenging task in the energy industry. In this paper, novel kernel-free $$\nu $$ -support vector regression model proposed for forecasting. The produces reduced quadratic surface nonlinear regression. A feature weighting strategy adopted to estimate relevance of features history. To reduce effects outliers history, weight assigned represent relative importance each data point. Some computational experiments are conducted some public benchmark sets show superior performance over widely used models. results extensive from Global Energy Forecasting Competition 2012 and ISO New England demonstrate better average accuracy model.