Forecasting Offshore RMB Exchange Rate Based On A Hybrid Model PCA-MCS-SVR-WOA

With the internationalization of RMB,the offshore RMB developing rapidly,which generating an profound impact on domestic RMB and China's economy.Additionally,with the exchange rate trading becoming popular,the exchange rate risk becomes an important problem.Therefore,forecasting offshore RMB exchange rate is important.At present,forecasting exchange rate is mostly based on its self-regulation by a hybrid model.This paper,forecasting the offshore RMB exchange rate based on the regulation of offshore and onshore RMB exchange rate by a hybrid model PCA-MCS-SVR-WOA which including Principal Component Analysis(PCA),MCS(the Model Confidence Set),three form of Support Vector Regressions(named as SVR),Whale Optimization Algorithm(WOA).Firstly,applying Copula method for correlation analysis between offshore and onshore RMB exchange rate.Secondly,selecting the input of SVR by PCA-MCS,which guarantees the input of SVR is irrelevance and optimality.Finally,adapting three form of SVR model,namely FSVR,vSVR,wSVR with the Whale Optimization Algorithm(WOA)to predict respectively,which also can verify the validity of PCA-MCS method,and comparing its results to determine the final predicting result.Using USD/CNH for empirical analysis,and the result of which show that the up-tail correlation between offshore and onshore RMB exchange rate,and the introduction of onshore RMB exchange rate is also help for improving the prediction ability,which means the onshore RMB exchange rate has a certain influence on the prediction of offshore RMB exchange rate;PCA-MCS method can improve the prediction ability obviously,and comparing with the benchmark model,the main model in this paper is superior.