Forecasting the renminbi exchange rate with an EMD-based neural network

This chapter looks at account two kinds of current exchange rates for the US dollar (USD) to the RMB. It focuses on applying artificial neural network (ANN) to forecast the renminbi (RMB) exchange rate with different forecasting horizons of 1, 5, 10, 20 and 30 days. The ANN is a computational model widely used in computer science and other research disciplines. It is based on a large collection of simple neural units and inspired by biological neural networks. Consider the two main criteria in the empirical experiments: the mean squared error (MSE) and the directional statistic. The chapter discusses an application used for designing trading strategy on onshore exchange market (CNY). With the continuing development of RMB internationalization, more policymakers, individual investors, financial institutions, and emultinational corporations are participating in RMB currency trading. The chapter considers Empirical mode decomposition (EMD), an adaptive time series decomposition technique that first proposed by Huang et al.