The extreme value distribution theory is an important tool in study of extreme value phenomena. This paper introduces univariate extreme value distribution models and multivariate ones. Besides, threshold models are considered due to the disadvantages of these block models. After study, we come to some valuable conclusions.These are the primary contents of this paper:First, we observe that the extreme value phenomena exist widely around us, which plays a part in the balance of nature and the development of our society. The research of these phenomena can help to grasp the laws and bring down the losses in difficulties.Second, univariate extreme value distribution models are used to analyze annual maximum rainfall data in Xuzhou. They can mimic the volatility behavior. The study results are fitted together with the owned conclusions.Third, modeling only block maxima is wasteful in extreme analysis if other data on extremes are available. Though the r largest order statistic model is a better alternative, it is unusual to have data of this form and threshold models are set up.Fourth, multivariate extreme value distribution models are introduced to analyze multivariate extreme values. Limit laws for multivariate extremes and the parameter estimates are studied, especially bivariate ones.
|