Exchange Rate Based On BP Neural Network Intelligent Forecasting System Development And Design

Since the central bank announced that the RMB exchange rate will implement the"The market supply and demand with reference to a basket of currencies, managedfloating exchange rate regime". Since all kinds of multi-currency trading andsettlement will be subject to varying degrees exchange rate fluctuations, the risk ofexchange rate fluctuations by how accurately predict the direction and extent ofchanges in foreign currency exchange rates is the basis of risk mana gement. Therefore,the exchange rate is of great significance. They can enhance their ability for exchangerate in the floating exchange rate predicted by the exchange rate.The exchange rate forecast fields filled with various non-model approach to theuse of artificial neural network model most widely used. Study of neural networks asa starting point, with its strong self-learning, association, recognition, and otherfunctions, to collect data to train the network, with its mathematical model to predictthe algorithm. BP neural network is one of the most representative and by the use of afairly broad front back propagation algorithm to train multilayer feed forwardnetworks. After training by studying historical data, which in turn provides a degreeah fit within the exchange rate forecast of samples. Improve the accuracy ofprediction, after MetaTrade4used in for ex-reading platform, and can be derived andused to predict the effect of a real environment.To establish intelligent forecasting system for the study, reviewed currentresearch methods to predict the exchange rate, to build intelligent forecasting systemto make some humble opinion: First, based on BP neural network predictive modelingtraining historical data; second, using BP neural network technical trading strategiesof exchange rate changes in the trend analysis and forecasting, the effect is significant;third, for exchange rate forecasting system needs analysis and system design; Finally,BP technology forecasting exchange rate trends Mat Lab neural network methodimplemented in software, and MetaQuoteSLanguage4with the current locale of themore common platform MetaTrade4contained discuss how to build (Expert Advisor.EA).