Selection of the Appropriate Lag Structure of Foreign Exchange Rates Forecasting Based on Autocorrelation Coefficient

We propose a new criterion, called autocorrelation coefficient criterion (ACC) to select the appropriate lag structure of foreign exchange rates forecasting with neural networks, and design the corresponding algorithm. The criterion and algorithm are data-driven in that there is no prior assumption about the models for time series under study. We conduct the experiments to compare the prediction performance of the neural networks based on the different lag structures by using the different criterions. The experiment results show that ACC performs best in selecting the appropriate lag structure for foreign exchange rates forecasting with neural networks.