RESEARCHING MODEL FORECAST THE CENTER EXCHANGE RATE USD/VND BY THE ANALYTICAL TECHNIQUES OF TIME SERIES BOX-JENKINS ARIMA

Abstract

The researcher builds and selects the model that best predicts the average exchange rate for USD/VND. The method is implemented using the analysis technique of time series Box-Jankins ARIMA (autoregressive integrated moving average) with the average center exchange rate data for period (monthly) from year 2005 to 2016 (2005M01 - 2016M12). The research data is collected by the author on the International Monetary Fund's website. The main objective of this report is apply the ARIMA model as well as analytical techniques to construct a model forecast the monthly exchange rate (average exchange rate) of USD/VND. In addition, the study also aims to suggest methods and analytical techniques of time series to help investors and enterprises can develop into the models forecast daily exchange rate or buying exchange rate, selling exchange rate of financial institutions. The fit of the model as well as the accuracy of prediction were assessed through the informations: Normalize BIC, Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE) And Root Mean Squared Error (RMSE).