<![CDATA[In analysis of time series, variance error always be assumed constant. Nevertheless, there are some financial economic data such as exchange rate data that variance error has not constant. This research describes an alternative model in analysis of time series which allowed variance error as an autoregressive process that recognized by GARCH (Generalized Autoregressive Conditional Heteroscedastic). GARCH model was used for modeling time series with residual condition variation according to the time. Estimated of parameter GARCH model can be used with MLE (Maximum Likelihood Estimation) method. At the end of this thesis is represented the variance model of exchange value of Rupiah (IDR) to American Dollar (USD). This research was based on daily data from January 2002 to December 2005. The result of the research show that rate of exchange value IDR to USD satisfied the assumption from GARCH model and GARCH (2,2) model is the most appropriate for the time series data. ]]>
Copyrights © 2007
