The Principle And Application Of Grey Least Squares Support Vector Machine

In consideration of the weakness of unsatisfactory forecast result that the traditional grey model for concussive data series with bigger dispersion degree, this paper adopt a method to weaken the stochastic volatility of energy consumption data, the method is to produce a sequence combined index acceleration with geometric mean, and then make grey prediction after converting it to smooth sequence which is fit for the grey model.Combine it with the least squares support vector machine, providing a GM(1,1)-LSSVM prediction model with a modified residual. We use the model to predict the gross of coal energy consumption of Liaoning province from 1996 to 2009. Experimental results show that the prediction accuracy of the modified model is better than that of the grey model and Least squares support vector machine model respectively, so it can be used as a new method in the prediction of energy consumption.