返回题库

判定正态分布:Y=XZY=XZY+ZY+Z

判定正态分布

专题
General / 综合
难度
L4

题目详情

Let XX and ZZ be independent random variables with XX a discrete random variable given by P(X=1)=P(X=1)=1/2P(X = 1) = P(X = - 1) = 1 / 2 and ZZ a standard normal random variable. Let Y=XZY = XZ . Is YY a standard normal variable? Is Y+ZY + Z normal? Does (Y,Z)(Y,Z) have joint normal density?

解析

X=±1X=\pm1 等概率,ZN(0,1)Z\sim N(0,1) 且独立,Y=XZY=XZ

  • YYZZ 同分布: P(Yt)=12P(Zt)+12P(Zt)=P(Zt).\mathbb{P}(Y\le t)=\tfrac12\mathbb{P}(Z\le t)+\tfrac12\mathbb{P}(-Z\le t)=\mathbb{P}(Z\le t). 所以 YN(0,1)Y\sim N(0,1)
  • Y+ZY+Z 不正态,因为 P(Y+Z=0)=P(X=1)=12\mathbb{P}(Y+Z=0)=\mathbb{P}(X=-1)=\tfrac12,有原子质量。
  • 因为联合正态下任意线性组合仍正态,而 Y+ZY+Z 非正态,所以 (Y,Z)(Y,Z) 不是联合正态。