| Summary: | 碩士 === 國立雲林科技大學 === 財務金融系碩士班 === 95 === This essay is going to have research on international gold spot price. We applied several time series models, such as one equation models (ARIMA Models、Regression with ARIMA error Models), simultaneous equation models (Vector Autoregressive Models (VAR)、Error Correction Models (ECM)) to evaluate the gold spot price. The data used in the analysis include information on the gold spot price、exchange rate、interest rate、crude oil price and consumer price index. The samples selected are mainly from two distinguished group, one is for building the gold spot price models from January 1986 to December 2004; the other is for forecasting purposes from January 2005 to December 2006. The forecasting performance is measured by comparing RMSE and MAD of each model. The important results are shown as follows：
(1) In one equation models, the result show regression ARIMA error models are better than ARIMA models. In Multi-equation models, the ECM models perform much better than VAR models.
(2) In addition, the ECM models also outperform than others models.
|
|---|
