Feature selection and parameter optimization of support vector regression for electric load forecasting

Forecasting of future electricity demand has become a promising issue for the electric power industry. Since many factors affect electric load data, machine learning methods are useful for electric load forecasting (ELF). On the one hand, it is important to determine the irrelevant factors as a preprocessing step for ELF. On the other hand, the performance of machine learning models depends heavily on the choice of its parameters. These problems are known respectively as feature selection and model selection problems. In this paper, we use the support vector regression (SVR) model for ELF. Our contribution consists of investigating the use the particle swarm optimization for both feature selection and model selection problems. Experimental results on two widely used electric load dataset show that our proposed hybrid method for feature selection and parameter optimization of SVR can achieve better results when compared with the classical SVR model while using feature selection and without using it.