汇率预测的理论与方法及其最新进展

总结了众多国内外学者在汇率预测方面的理论及其研究方法。通过对传统的统计方法与非参数方法的比较分析,得出结论:大多数传统的时间序列模型是线性的,不能抓住非线性时间序列数据的内在特征。而相对于传统的预测模型而言,非参数方法能发现观察结果和输入数据的关系,不需要事先确定模型,其拟合结果能更好的捕捉汇率的动态特征与走势。

Theory and Method of Exchange Rate Forecasting and the Latest Progress

This paper summarizes most of the theory and method of exchange rate forecasting held by many global famous scholars. After comparing the traditional statistic methods with the non-parameter methods, we draw a conclusion that: most traditional time series models are linear and cannot catch the inner character of those non-linear time series. However, non-parameter methods can examine the relations between the results with its input data without defining a model beforehand. The most important is its simulation results can catch the exchange rate dynamics and characters more efficiently.