The Research And Analysis Of Exchange Rate Forecasting Model Based On GA-SVR

As an important part of international finance, exchange rate is always the hot topic for discussion and research. Since July 7, 2005, China has implemented a new exchange rate system that adjusts the exchange rate on the basis of the supply and demand of the market as well as a referenced basket of currencies. It is supervisory floating system. It is assumed that this new exchange rate system may lead to a day-by-day increase of frequent and instable fluctuation of exchange rate, which has negative influences on our financial system, such as increasing the exchange rate risk, pricking up speculation, increasing the instability of financial market and so on. Therefore, it is indispensable to establish a prompt and accurate exchange rate forecast model to capture the fluctuation of RMB/USD exchange rate in certain future period.This dissertation first discusses the development of exchange rate forecast methods, together with the applications of Support Vector Machines (SVM) in forecast and its limitations. Then the thesis introduces the theory of SVM and Genetic Algorithms (GA) by chiefly analyzing the limitations of SVM and the feasibility of optimizing the SVM using GA. Subsequently, it proposes a scheme to optimize SVM using GA encoded with real value. On basis of the analyzed classic theories of foreign exchange rate and the characters of RMB, in the fourth part of the thesis, 10 macro economic indexes of China and USA are presented as the factors of influencing RMB/USD exchange rate. Then a GA-SVR exchange rate forecast model based on these 10 structured variables is established. To apply the model, data from Jan. 1999 to Oct. 2005 are quoted to make empirical research on exchange rate. The result indicates that the forecast precision of GA-SVR is better than that of standard SVR. The MAPE of it is only 0.41777%, showing the forecast values fit the actual ones very well. By referring to the problems which emerge in the above model, like the prematurely problem of GA and so on, the last part of the thesis applies the Adaptive Genetic Algorithms (AGA) and the method of combining structured variables to optimize and adjust the model separately. The evaluating indicators show that the effect of optimization and adjusting is very good. They not only solve the problems mentioned above, but also improve the forecast precision further.