0
0

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?

More than 1 year has passed since last update.

条件付き分布の2つの導出

Last updated at Posted at 2022-06-04

条件付き分布を2つの方法で示したい

統計検定準1級に出題された問題をまとめました.
1つ目の方法より,2つ目の方法の方が自分的には覚えやすかったです.

示すこと

\left(\begin{array}{l}
X \\
Y
\end{array}\right) \sim N\left(\left(\begin{array}{l}
\mu_{x} \\
\mu_{y}
\end{array}\right),\left(\begin{array}{cc}
\sigma_{x}^{2} & \rho \sigma_{x} \sigma_{y} \\
\rho \sigma_{x} \sigma_{y} & \sigma_{y}^{2}
\end{array}\right)\right)

のとき,$X=x$を与えたときの$Y$の条件付き分布は,正規分布

N\left(\frac{\mu_{y}+\rho \sigma_{y}\left(x-\mu_{x}\right)}{\sigma_{x}},\left(1-\rho^{2}\right) \sigma_{u}^{2}\right)

で与えられる.このことを2通りの方法で示す.

証明1

X, Yをそれぞれ標準化して,

X^{*}=\frac{X-\mu_{x}}{\sigma_{x}}, \quad Y^{*}=\frac{Y-\mu_{y}}{\sigma_{y}} 

とおけば,

\left(\begin{array}{l}
X^{*} \\
Y^{*}
\end{array}\right) \sim N\left(\left(\begin{array}{l}
0 \\
0
\end{array}\right),\left(\begin{array}{ll}
1 & \rho \\
\rho & 1
\end{array}\right)\right)

であるので,

Z=Y^{*}-\rho X^{*}

の期待値と,分散を考える.
期待値は,

E(Z)=E(Y^{*}-\rho X^{*})=E(Y^{*})-\rho E(X^{*})=0
\begin{aligned}
V(Z)&=V(Y^{*}-\rho X^{*})=V(Y^{*})-2\rho Cov(X^{*}, Y^{*}) + V(Y^{*}) \\
 &=1-2\rho^2 + \rho^2=1-\rho^2
\end{aligned}

となる.
つまり, $Z \sim N\left(0,1-\rho^{2}\right)$となる.したがって,$X=x$を与えたときのYの平均と分散は,それぞれ

\begin{aligned}
E(Y \mid X=x) &=\mu_{y}+\sigma_{y} E\left(Y^{*} \mid X=x\right) \\
&=\mu_{y}+\sigma_{y} E\left(Z+\rho \frac{x-\mu_{x}}{\sigma_{x}}\right) \\
&=\mu_{y}+\rho \sigma_{y} \frac{x-\mu_{x}}{\sigma_{x}} \\
V(Y \mid X=x) &=\sigma_{y}^{2} V\left(Y^{*} \mid X=x\right) \\
&=\sigma_{y}^{2} V(X) \\
&=\sigma_{y}^{2}\left(1-\rho^{2}\right)
\end{aligned}

となる.

証明2

現代数理統計学(竹村先生)のp59, 60の結果を用いる.
$Y=\left(Y_{(1)}, Y_{(2)}\right) $と分割し $ Y_{(1)}=\left(Y_{1}, \ldots, Y_{h}\right)$の周辺分布を考える.

\mu=\left(\begin{array}{l}
\mu_{1} \\
\mu_{2}
\end{array}\right), \quad \Sigma=\left(\begin{array}{cc}
\Sigma_{11} & \Sigma_{12} \\
\Sigma_{21} & \Sigma_{22}
\end{array}\right)

このとき,
$Y_{1}$が与えられたときの$Y_{2}$の条件付き分布を考えると,

$$
Y_{2} \mid Y_{1}=y_{1} \sim \mathrm{N}\left(\mu_{2}+\Sigma_{21} \Sigma_{11}^{-1}\left(y_{1}-\mu_{1}\right), \Sigma_{22}-\Sigma_{21} \Sigma_{11}^{-1} \Sigma_{12}\right)
$$

となる.
この事実を用いると,先ほど示したかったことが証明できる.

0
0
0

Register as a new user and use Qiita more conveniently

  1. You get articles that match your needs
  2. You can efficiently read back useful information
  3. You can use dark theme
What you can do with signing up
0
0

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?