2
1

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.

4.1(標準) コーシー・シュワルツの不等式,共分散

Last updated at Posted at 2021-08-26

方針

コーシー・シュバルツの不等式:

\mathbb{E}_{X\sim f_X}[(X-\mu_X)^2]\mathbb{E}_{Y\sim f_Y}[(Y-\mu_Y)^2]\geq\mathbb{E}_{X\sim f_X,Y\sim f_Y}[(X-\mu_X)(Y-\mu_Y)]^2

を証明する.これより,相関係数:

Corr(X,Y)=\frac{\mathbb{E}_{X\sim f_X,Y\sim f_Y}[(X-\mu_X)(Y-\mu_Y)]}{\sqrt{\mathbb{E}_{X\sim f_X}[(X-\mu_X)^2]}\sqrt{\mathbb{E}_{Y\sim f_Y}[(Y-\mu_Y)^2]}}

について,

|Corr(X,Y)|\leq 1

が成り立つ.問題では,tについての二次関数:

h(t)= \mathbb{E}_{X\sim f_X,Y\sim f_Y}[\{(X-\mu_X)-t(Y-\mu_Y)\}^2]

の判別式を求めることでコーシー・シュワルツの不等式を証明する.

答案

tについての二次関数:

\begin{align}
h(t)&= \mathbb{E}_{X\sim f_X,Y\sim f_Y}[\{(X-\mu_X)-t(Y-\mu_Y)\}^2]\\
&= t^2\mathbb{E}_{Y\sim f_Y}[(Y-\mu_Y)^2]-2t\mathbb{E}_{X\sim f_X,Y\sim f_Y}[(X-\mu_X)(Y-\mu_Y)]+\mathbb{E}_{X\sim f_X}[(X-\mu_X)^2]
\end{align}

について,定義より,任意のtに対し,

h(t)\geq 0

が成り立つので,二次方程式h(t)=0の判別式をDとおくと,

D=4(\mathbb{E}_{X\sim f_X,Y\sim f_Y}[(X-\mu_X)(Y-\mu_Y)])^2-4\mathbb{E}_{X\sim f_X}[(X-\mu_X)^2]\mathbb{E}[(Y-\mu_Y)^2]\leq 0

なので,コーシー・シュワルツの不等式が証明された.

次に,コーシー・シュワルツの不等式の等号成立条件を考える.h(t)を平方完成して,

\begin{align}
h(t)&= t^xVar(Y)-2tCov(X,Y)+Var(X)\\
&= Var(Y)\{t^2-2t\frac{Cov(X,Y)}{Var(Y)}\}+Var(X)\\
&= Var(Y)(t-\frac{(Cov(X,Y)}{Var(Y)})^2-\frac{\{Cov(X,Y)\}^2}{Var(Y)}+Var(X)
\end{align}

より,

-\frac{\{Cov(X,Y)\}^2}{Var(Y)}+Var(X)=0

の時h(t)=0なので,

Cov(X,Y)=\pm\sqrt{Var(X)}\sqrt{Var(Y)}

の時等号成立.

参考文献

2
1
2

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
2
1

Delete article

Deleted articles cannot be recovered.

Draft of this article would be also deleted.

Are you sure you want to delete this article?