Hybrid Genetic Algorithm and Support Vector Regression Performance in CNY Exchange Rate Prediction

摘要

In this paper, we predict CNY exchange rates in terms of hybrid genetic algorithm and support vector regression with a range of kernel functions. A BP neural network model is benchmarked with the hybrid model and then the hybrid genetic algorithm and support vector regression model are used to examine the accuracy of CNY exchange rate prediction. The intuitive and statistical performances of the hybrid model with linear, radical basis, polynomial and sigmoid functions are presented and analyzed by using the exchange rate data of USD/CNY, EUR/CNY and CNY/JPY. The empirical results show that the hybrid model is effective for studying the CNY exchange rate prediction.