Prediction For Futures Prices Based On Predicting Technique Of Nonlinear Integration

With the development of socialist market economy, the status of futuresmarkets in the social economic is becoming more and more important.Especially, embodied in the stock market risk aversion, reasonable allocationof social resources and stable financial market order. Have a accurate grasp offutures market becomes very important. If these information want to manifestto specific quantitative index, we must have a accurate prediction of futuresmarkets. Due to the complexity of the internal structure of the futures market,and influence factors of diversity. Existing prediction method in a certain extent,there is still room for improvement. Based on above analysis, reasonable andaccurate forecasting of futures price movements becomes a focusThis paper aims to establish a forecasting model of nonlinear integrated topredict futures price with more accurate. Trying to explore a more effectivenonlinear integration method to integrate more and more effective information.These information can contribute to improve the futures price predictionprecision. The author uses this technology to forecast the futures price.(Example for Zhengzhou Commodity Exchange), also the first try use thistechnology to forecast futures price of Zhengzhou Commodity Exchange.Firstly, the article from the background knowledge of futures market tointroduces the relevant knowledge, including the concept and characteristic offutures, trading in futures market both at home and abroad, futures species.Meanwhile analyzes the futures price prediction methods. Then we select thesingle model which included in integration model of nonlinear. The article useintegration model of nonlinear as its key point to integrate ARIMA model、BP-ANN model、VAR model.Secondly, In order to make the integration model of nonlinear containsmore accurate information; the prediction of single model must has enoughforecasting accuracy. So, the key point of ARIMA model is to determine thevalue of p and q.The article summaries the research of previous and propose thatthe way of scientific is applied to determine the optimal combination offorecasts on the confirmation of the p value and q value.Because the articledeal with time-series data,how to choose the perfect point of input layer of BP-ANN is a difficult question.In order to solve this problem,the articlethrough delayed entry of ARIMA model to choose the best point of input layerof BP-ANN. Using this way to solve question,We avoid subjectivity by singleperson.Finally, this paper uses Zhengzhou Commodity Exchange futures data forempirical studies, predicts the prices of Cotton futures, sugar futures and PATfutures respectively. The empirical results show that the integration model ofnonlinear can achieve satisfactory effect.