Utilizing IABC and Time Series Model in Investigating the Influence of Adding Monitoring Indicator for Foreign Exchange Rate Forecasting

Abstract

This work focuses on the NTD/USD exchange rate and the Monitoring Indicator in years of 2006 to 2010 to forecast the foreign exchange rate via Time-series models including GARCH (1,1) and EGARCH (1,1), and a computational intelligence model called IABC. In order to compare the rate forecasting ability of these models, the MAPE is consecutively applied as the evaluating criterion after the forecasting process. The experimental results indicate that it is effective to enhance the ability of foreign exchange rate forecasting by adding the Monitoring Indicator as a new reference variable in the IABC model. Based on the experimental results, we find that IABC is the most effective one to forecast the foreign exchange rate. Nevertheless, when IABC is suffered from the local optimum in the solution space, the forecasting ability would present a significant drop.