金融危机背景下的人民币汇率预测

在为金融危机期间人民币汇率的波动提供一种有效的预测方法.在利用替代数据方法检验和判别汇率系统具有非线性结构的基础上,识别了各具体汇率序列的最优滞后期组合,并分别采用了多层感知机(MLP)和层反馈网络(RNN2)结构构建同质神经网络模型,从三个方面对比分析了模型群在不同参数条件下的预测效果. 研究发现,根据不同序列的具体特征,各神经网络模型在不同自由度下的4个预测期限内的预测性能存在较明显的差异.同时,包含层反馈过程的RNN2模型在描述与预测人民币汇率的波动方面表现出很强的能力.此外, 还分析并解释了产生上述结果的原因,并为4种人民币汇率波动序列甄选出了相应的最优预测模型.

RMB exchange rate forecasting in the context of the financial crisis

This article offered an effective method to forecast the RMB exchange rate volatility during the financial crisis period. Based on the inspecting and discriminating the nonlinearities structure of the exchange rate system by using the method of surrogate data, the optimal lag periods for each specific exchange rate series was computed, then we adopt both the structure of multi-layer perceptrons (MLP) and recurrent neural networks (RNN2) to build the homogenous ANN model. We compared the forecast results of all the ANN models with different parameters from three respects. It showed that, according to the specific characteristics of each exchange rate series, the forecast performance of ANN models with different freedom degrees have obvious differences in the four forecast periods. And the RNN2 model, which contains layer feedback process, has showed great ability to explain and forecast the behavior of the volatility of the RMB exchange rates. The study has explained the reasons for these results, and chooses the optimal forecasting model for each RMB exchange rate volatility series.