Export Forecasting Based On Fuzzy Neural Network

Since the reform and opening up, China's foreign trade has made great achievements. Since 1991, China's opening up gradually deepened, provinces, municipalities and autonomous regions, the autonomy of import and export trade, the gradual liberalization of foreign trade enterprises make their own decisions, Right and Interest in the combination of a great mobilization of the regional initiative to develop foreign trade. Especially in 2001 China's accession to the WTO, China's foreign trade has a broader space for development. Moreover, because of information technology, particularly the rapid development of database technology, we can save a lot of foreign trade information. We can analyze and forecast exports information. Correct and reasonable export predictions not only provide decision support for the relevant departments to formulate effective policies, but also rightly organize and guide enterprises to export behavior, so as to optimize and upgrade the purpose of the export product mix. However, the current research on the foreign information has shortcomings in domestic and international, one existing on the export trade of the index system can not reflect the social, economic and policy changes and other factors on the impact of foreign trade, needs to expand foreign trade index system; Second, the current import and export the results less predictable.This paper analyzes the significance and methods of foreign trade prediction, and introduces export forecast factors, characteristics and present conditions. By analyzing the domestic and foreign research of export forecast forecasting to building the forecasting index system. Elaborates of artificial neural networks and fuzzy theory concepts and principles, analyzes their advantages and disadvantages, also about complementarities between them. Describing the type of fuzzy neural network and training algorithm, neural network and fuzzy theory combined forecasting model. Finally, this paper discusses characteristics of Chongqing exports, factors taken into account the impact of forecasting, a fuzzy neural network model is applied to Chongqing's foreign trade exports forecasting. Discuss the whole process about the fuzzy neural network prediction model building. Analysis Change Interval of the Chongqing exports forecasting to obtain a more satisfactory prediction results. Compared with the ARIMA, regression analysis, gray model, support vector machines and neural network prediction methods. The results show that the fuzzy neural network is superior to other forecasting methods, which can predict changes in foreign trade and exports of Chongqing, the government can provide a more reliable basis for decision-making.