A New Prediction Method for Short-Term Forecast

Abstract

This study introduced a new method for the short-term forecast theat actually is a hybrid model, combing the grey model and the cumulative least squared linear model, with the special feature of automatically adjusting the overestimated or underestimated predicted value around the points having extreme value. The verification of this study is also tested successfully in two experiments. This demonstrated theat the proposed method has the best accuracy of predicted value among four short-term forecasting models discussed in this study.