A Comparative Study of Univariate Forecasting Methods Based on the Exchange Rate between New Taiwan Dollar and U.S Dollar

• Main Authors: LU,HUI-FANG, 盧慧芳
• Other Authors: Ching-Wu Chu
• Format: Others
• Language: zh-TW
• Published: 2013
• Online Access: http://ndltd.ncl.edu.tw/handle/90298265894660341464

碩士 === 國立臺灣海洋大學 === 航運管理學系 === 101 === As the liberalization of global trade, international trade is increasing. The floating exchange rate mechanism results in the uncertainty on the exchange rate and produces a huge variable for the international trade and investments. If we have a good grasp of trend variations of the exchange rate and make an apporiate hedge against fluctuations, it can effectively reduce the fluctuation risk of exchange rate. The purpose of this study is to provide a more accurate prediction model on the forecasting based on NT - US exchange rate for monthly data. Five different univariate methods, namely the Classical Decomposition Model, the Regression Model with Seasonal Dummy Variables, the Trigonometric Model, the Grey Forecast, the Hybrid Grey model , have been used. The contribution of this research is to compare the forecasting results of the five univariate methods based on commonly used evaluation criteria, MAE, MAPE and RMSE. We found that, the Regression Model with Seasonal Dummy Variables model is a reliable prediction method for forecasting NT - US exchange rate. The results of this work can be a helpful to reference the forecasting NT - US exchange rate.