摘要：The exchange rates of four currencies in South Asia depend on the Foreign exchange market amongthose countries. In most cases, the best fitting distribution of exchange rates are found to follow the generalizedlogistic distribution. Now, we seek for the best distribution to describe the extreme daily exchange rate from1993-2013 for 4 countries in Southeast Asia using the classical and Bayesian methods. We use the maximumlikelihood method for classical approach while Generalized Extreme Value (GEV) distribution and generalizedlogistic distribution for Bayesian approach. The estimate of the three-parameters of extreme values distributionhave been compared using goodness-of-fit test. The tests show us that there is no significant difference ofparameter estimated between these two distributions. Furthermore, the best distribution for fitting the modelof these exchange rates is GEV.
