两个正态总体方差双边齐性检验的优化  

作　　者：李婉 张蓓蓓 武丹 

机构地区：[1]南京医科大学康达学院理学部,江苏连云港222000

出　　处：《科技创新导报》2020年第16期210-211,213,共3页Science and Technology Innovation Herald

摘　　要：在对两个正态总体的方差进行双边齐性检验时,由于检验统计量服从分布,而这个分布的密度函数关于它的峰值不是对称的,因此传统的方法给出的临界值不是最佳的。本文提供了一种方差双边齐性检验的优化方法,使得置信度(第一类错误)在给定数值的前提下,第二类错误的发生降低到最小。并在此优化方法的基础上,本文借助软件给出了部分最佳临界值表,以供假设检验使用。When performing the two-side homogeneity test on the variance of two normal populations,since the test statistic obeys the F distribution,and the density function of this distribution is not symmetric about its peak value,the threshold value given by the traditional method is not the best.This paper provides an optimization method for two-side homogeneity test of variance,so that the occurrence of typeⅡerror is minimized,under the premise of a given value of the degree of confidence(typeⅠerror).On the basis of this optimization method and with the help of software,this paper gives some best critical value tables for hypothesis testing.

关 键 词：两个正态总体 方差齐性检验 最佳临界值表 

分 类 号：R195[医药卫生—卫生统计学]