两样本相等性的面积统计量检验  

Testing the Equality of Two Distributions by Area Statistics

作　　者：卢梦霞[1] 凡美金 

机构地区：[1]周口师范学院数学系,河南周口466001 [2]中国矿业大学理学院,北京100083

出　　处：《河南师范大学学报（自然科学版）》2014年第2期29-32,共4页Journal of Henan Normal University(Natural Science Edition)

基　　金：国家自然科学基金(11001267;11171001);中央高校基本科研业务费(2009QS02)

摘　　要：分布相同的非参数假设检验是统计学者一直关心的问题之一.关于面积统计量检验总体相等性,提出新的面积统计量TA1,TA2,TA3,TA4及相应的Bootstrap统计量TA1^,TA2^,TA3^,TA4^,并利用投影寻踪(PP)技术扩展到多维情况.证明了它们在原假设下有相同的极限分布,并进行了大量的模拟试验,验证选择上述统计量的合理性.Testing the equality of two distributions with nonparametric has long been an interesting issue in statistical inference. In this paper, we shall use the bootstrap and area statistic test TA1, Ta2, TAa, Ta4 and corresponding Bootstrapping test TA1 ,Ta2,Taa,TA4. And we use projection pursuit to expand this program from one-dimension to several-dimensions. In the case that the assumption is hold, some results on the limiting distributions of the statistics are obtained. Several simulation experiments are performed, and we get satisfied results.

关 键 词：BOOTSTRAP方法 面积统计量 K-S统计量 投影寻踪 

分 类 号：O213.2[理学—概率论与数理统计]