Forecasting Method of Multivariate Time Series Applied in Short-term Electrical Load Forecasting

Univariate time series is extended to multivariate time series in order to improve the forecasting accuracy and to fully make use of superiority of chaotic short-term forecasting.At the same time,the phase space of multivariate time series is reconstructed,and embedding dimensions and delay time are calculated.Electrical load is forecasted by RBF net.Forecasting result shows the method is more effective than univariate time series.