On Forecasting the Price of the Futures of Foreign Exchange.

[[abstract]]本文以ARIMA模型與GARCH模型以及移動迴歸法，研究五種外匯期貨，並以絕對平均誤差、均方誤差 、絕對平均百分比誤差做為預測績效之衡量標準。結果顯示:五種外匯期貨之預測績效並無顯著差異。預測得到的每日報酬率還原成每日期貨價格，即為預期之每日外匯期貨價格。此一預測值，可運用到財務投資管理、公司決策，以及投資參考依據。[[abstract]]This paper explores the exchange Future of five currencies with ARIMA, GARCH models, and rolling regression method . The MSE, RMSE and MAPE are used to measure the performance of forecasting. The result shows that the forecasting performance of the exchange future of five currencies are not significantly different. The forecasted daily returns are turned the daily price of exchange future into be the predicted price of exchange future, which could be used as the guideline for the managing of financial portfolio and corporations' decision-making