节点文献
二项-广义Pareto复合极值分布模型的统计推断
Statistical Inference for Binomial-generalized Pareto Compound Extreme Value Distribution Model
【摘要】 极值理论主要研究小概率、大影响的极端事件.当前,复合极值分布已经广泛应用于水文、气象、地震、保险、金融等领域.本文以极值类型定理和PBDH定理为理论依据,构建了二项-广义Pareto复合极值分布模型;使用概率加权矩方法,对所建立的复合模型推导参数估计式;利用计算机模拟,得到了Kolmogorov-Smirnov(简称KS)检验统计量的临界值.
【Abstract】 Extreme value theory is mainly the study on extreme events of small probability & major impact.At present,the compound extreme value distribution has been widely used in hydrology,meteorology,earthquake,insurance,finance and other fields.In this paper,we establish binomial-generalized Pareto compound extreme value distribution model based on extreme value type theorem and PBDH theorem,derive parameter estimation of the established compound model by probability weighted moments,get critical values of Kolmogorov-Smirnov test statistic.
【关键词】 广义Pareto分布;
二项分布;
概率加权矩;
Anderson-Darling(AD)检验;
KS检验;
次序统计量;
随机模拟;
【Key words】 generalized Pareto distribution; Binomial distribution; probability weighted moment; Anderson-Darling test; Kolmogorov-Smirnov test; order statistic; random simulation;
【Key words】 generalized Pareto distribution; Binomial distribution; probability weighted moment; Anderson-Darling test; Kolmogorov-Smirnov test; order statistic; random simulation;
【基金】 2010年北京市教委科技面上(KM201010005006);北京工业大学人才强教深化计划211工程-服务北京优秀团队(00600054R0001)资助项目
- 【文献出处】 应用数学学报 ,Acta Mathematicae Applicatae Sinica , 编辑部邮箱 ,2012年03期
- 【分类号】O211.3
- 【被引频次】9
- 【下载频次】243