![]() | Only 14 pages are availabe for public view |
Abstract When natural disasters happen, their occurrences and frequency should be studied to find whether anything could have been done to prevent them or at least to be prepared for them. Thus the determination of the distribution of extremes of a certain phenomenon is very important to obtain good solutions to the devastating disasters. Extreme value theory has two approaches for dealing with extremes; the block maxima and the peaks over threshold (POT) approach. In this study the peaks over threshold approach is focused on. Since the generalized Pareto distribution (GPD) is considered the limiting distribution of exceedances under the POT approach, the estimation of its parameters is of great interest. There are many estimation methods for the parameters of the GPD. So, some of the estimators of the parameters of the GPD are compared by simulation. Then these estimates were used to estimate value at risk and conditional tail expectation as a tail risk measures. Also the importance of this study appears in the ability of application in different fields. So, at the end five different datasets (insurance, hydrology, wind speed, price of gold and stock market) were analyzed using the POT approach. And vital questions related to the datasets were answered using the POT approach. |