Modeling And Forecasting White Sugar Futures Price: Evidence From Zhengzhou Commodity Exchange

White Sugar Futures, which began trading on January6,2006on the Zhengzhou CommodityExchange (ZCE), is conducive to help China participate in international pricing, and improveChinese sugar industry competitiveness in the global market. The influence of internationalpricing power is the role of the price of sugar in the spot market and international trade. And,sugar future price formed under the influence of various factors, so it is difficult to measure andforecast. However this study will want to put lots of factors into forecasting model, and put thispoint as the main work in this study.By reviewing the domestic and international research about futures prices forecasting, most ofthem ignored the possibility of forecasting before they forecast the future price. This study willwork on the possibility of forecasting sugar futures price and then found that white sugar futureClosing price has distribution characteristics of heavy–tailed. Meanwhile after Hurst indextesting, white sugar futures Closing price exists non-linear structure. These points mean thatsugar future did not reach Weak-Form EMH in China, in other words it is possible to predict theprice of white sugar future.This study select14factors, which effect on the price of white sugar future, to establish a data-driven model (Artificial neural network(ANN)) to predict Closing price. And RBF network ismore working than BP network, this maybe the result from that the RBF can indefinitelyapproximation nonlinear mapping. At the same time the study use the Forecast Combination,which combine structure prediction model with non-structure prediction model. But the effect ofForecast Combination is worse than ANN.