Least squares support vector machine model optimized by particle swarm optimization for electricity price forecasting

The parameters of least squares support vector machine (LS-SVM) are optimized by using particle swarm optimization (PSO) algorithm, and a new model of electricity price forecasting is presented. In the proposed model, LS-SVM that has well generalization performance and quick operation ability is used for modeling for time series electricity price data. In order to avoid blindness and inaccuracy in the choice of the parameters of the LS-SVM, the k-fold cross-validation error is selected as the target value on which the parameters are chose based, and particle swarm optimization algorithm that has global optimization capability is used for choosing the parameters of the support vector machine. The historical data from PJM market is used in the case study to forecast the day-ahead system marginal price. The simulation research results show that the PSO algorithm can tune the parameters of the LS-SVM and the proposed model can improve the precision of electricity price forecasting effectively.