The Hybrid Prediction Model of CNY/USD Exchange Rate Based on Wavelet and Support Vector Regression

Since the implementation of the new mechanism of Renminbi exchange rate from 2005, the CNY/USD exchange rate fluctuation range has become more greater than before. Therefore, it is very important to control CNY/USD exchange rate risk via prediction. This paper is motivated by evidence that different prediction models can complement each other in approximating data sets, and presents a hybrid prediction model of support vector machines (SVMs) and discrete wavelet transform (DWT) to solve the exchange rate prediction problems. The presented model greatly improves the prediction performance of the individual SVMs models in prediction exchange rate. In the experiment, the performance of the hybrid prediction model is evaluated using the CNY/USD exchange rate market data. Experimental results indicate that the hybrid prediction model outperforms the individual SVMs models in terms of root mean square error (RMSE) metric. This hybrid prediction model yields better prediction result than the individual SVMs models.