| 520 |
|
|
|a Modelling and forecasting financial time series data has become the area of interest in financial world. However, the data exhibits certain stylized facts that must be handled by an appropriate models. Thus, this study was conducted to develop hybridization models between Autoregressive Integrated Moving Average (ARIMA) model and Generalized Autoregressive Conditional Heterocedasticity (GARCH) family model for daily exchange rate data. Later, the performance of modelling and forecasting for the best models among them will be compared. GARCH family models are divided into two categories which are symmetric (GARCH) and asymmetric (EGARCH) models. In this study, daily data of U.S. Dollar exchange rate against Malaysia exchange rate (USD/MYR) is used from the period of 1st November 2010 until 30th August 2016 collected from the Central Bank of Malaysia. The data are divided into two parts where 90% of the data is used as in-sample period taken from 1st November 2010 until 3rd February 2016. Meanwhile, for another 10% is used for the out-sample period taken from 4th February 2016 until 30th August 2016. EViews software and Microsoft Excel are used in this study to analyze the data. The performance of the hybrid models are evaluated using AIC, MAE, RMSE and MAPE. Results showed that, hybrid ARIMA-EGARCH model is the best model in modelling and forecasting daily exchange rate data compared to hybrid ARIMA-GARCH model.
|
