ARIMA and neural network prediction of foreign exchange reserves
Chunhua Shi, Huimin Wang, Fancheng Yin, RU Zheng-liang · 2011
This paper is about ARIMA and neural network prediction of the foreign exchange reserves of China. Both of unit-root nonstationarity and nonlinearity are tested. In the conclusion, we show that the predictive accuracy of neural networks outperforms ARIMA in terms of the MSE and MADE criteria.