Exchange Rate Analysis And Forecast Based On R Language And Neural Network

Exchange rate prediction has always been one of the research and attention of Chinese and foreign scholars.Proper and effective exchange rate forecasting can help countries formulate and adjust interest rate-related policies.In addition,in the context of economic globalization,exchange rate forecasting is also of great significance to financial institutions,foreign trade companies,and ordinary people involved in the exchange rate market.Therefore,correct and effective exchange rate fluctuations and trend predictions are undoubtedly a matter of vital importance to the financial industry.This paper selects the daily exchange rate data for the whole year of 2016,selects R software as the tool of data analysis,and uses the nonlinear combination forecasting method based on BP neural network to select two commonly used exchange rate single item forecasting methods to analyze the data.Then,BP neural network is used to nonlinearly combine them to establish an exchange rate prediction model.This paper successfully uses this model to make short-term predictions of exchange rate data,and the prediction results are better than single prediction model and linear combination prediction model,which has certain application prospects.However,this method does not perform well in long-term forecasting.The error is too large when predicting the exchange rate data in April 2017 and beyond,indicating that the application scenario is limited and still needs improvement.