【摘要】 旨在研究人民币汇率在相对较长时间内的趋势及更优的预测模型.基于时间序列分解原理将人民币兑美元汇率的时间序列分解为长期趋势及随机波动两部分,采用Logistic曲线拟合长期趋势,建立了包含长期趋势的复合预测模型.实证检验结果表明此预测模型的预测效果优于单一的ARIMA模型,根据分离的长期趋势,可以判定,人民币尚处于一个升值过程中,其未来的升值极限在5.5左右.更多还原

【Abstract】 This paper is aimed at the long-term trend of RMB exchange rate and better forecasting model.The time series of RMB to USD exchange rate was decomposed into the product of long-term trend and random fluctuations,Logistic Curve was adopted to characterize long-term trend,a new composite forecasting model including long-term trend was set up.The test result showed that this model is better than single ARIMA model.According to the long-term trend,it can be determined that RMB is still in the process of appreciation,the ultimate value in the future will be around 5.5.更多还原