TIME SERIES PREDICTION BASED ON SUPPORT VECTOR REGRESSION

This paper introduced the research condition and the basic theory of support vector regression (SVR).SVR is applied to forecast data of Box-Jenkins gas furnace ,and compared with other feedforward network-back-propagate (BP) neural network, self-adaptive expanding neural network. The result of simulation shows that BP neural network and self-adaptive expanding neural network are relatively close in predicting performance , and SVR is obviously superior to these two kinds of methods in predicting performance. The article analyses the reason why SVR is superior to the BP neural network and self-adaptive expanding neural network.