正态数据偏度峰度检验法
置信度:1-alpha
数据样本数:n
sig1=sqrt(6.0*(n-2)/(n+1)/(n+3));
sig2=sqrt(24.0*n*(n-2)*(n-3)/(n+1)/(n+1)/(n+3)/(n+5));
mu2=3.0-6.0/(n+1);
样本中心矩
B2=A2-A1*A1;
B3=A3-3.0*A2*A1+2.0*A1*A1*A1;
B4=A4-4.0*A3*A1+6.0*A2*A1*A1-3.0*A1*A1*A1*A1;
Ak为k阶样本矩 sum(xi^k)/n
样本偏度和峰度:
g1=B3/(B2)^(3/2)
g2=B4/B2/B2;
u1=g1/sig1;
u2=(g2-mu2)/sig2;
H0为真,n充分大时 u1、u2标准正态分布
拒绝域 fabs(u1)>=z(alpha/4.0) and fabs(u2)>=z(alpha/4.0);
要求n>=100;
时间: 2024-10-30 01:26:58