Research On Foreign Exchange Rate Forecast Based On Neural Network

Since the reform and opening up,with the rapid improvement of people's living standards,the development of the domestic financial market has gradually deepened,the foreign exchange market has been favored by many investors.At the same time,the foreign exchange risk also aroused people's attention and widespread concern.How to manage and avoid foreign exchange risk well is an important course for many foreign exchange investors.The foreign exchange market has the complexity,variability and high volatility characteristics,many factors influence the fluctuation of exchange rate varied,perplexing,is a typical nonlinear system,it is very difficult to make accurate predictions.Therefore,the prediction of foreign e xchange rate has become a hot issue in the field of foreign exchange risk management and data mining.According to the nonlinear model of BP neural network can approximate arbitrary and excellent characteristics to forecast short term,select the BP neural network to predict the yen / dollar and GBP / USD two on the currency,forecast above two pairs of 5 minutes,15 minutes and 30 minutes of the closing price,the highest price and the lowest price.And the Arima model is compared with the BP neural network model to illustrate the excellent performance of the BP network in the short-term prediction.The neural network function is constructed by MATLAB,and the corresponding delay factors are added to the network according to the characteristics of the time delay in the foreign exchange market.With the ratio of 8:2 to divide the training set and test set,the test set results,with input delay factor of the BP neural network can better predict the short-term exchange rate,and has certain reference value to the investment decisions of investors.