2
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 3 years have passed since last update.

2.11(標準) 確率母関数 (方針で確率母関数を解説)

Posted at

方針

前半のkと後半のkの意味が違うっぽい.略解見る限り,問題文は


f_X(x)=\frac{1}{2^{x+1}},\ \ x=0,1,...

が確率関数になることを示せ.この分布の確率母関数と積率母関数を求め,正の整数kに対して

\mathbb{E}_{X\sim f_X}[X(X-1)...(X-k+1)]

を求めよ.


の方がわかりやすいと思う.

この問題のメインは離散確率変数に対して定義される確率母関数.離散確率変数Xに対し,確率母関数G_Xとは,

G_X(s)=\mathbb{E}_{X\sim f_X}[s^X]

という,定数sについての関数である.これを微分し,s=1を代入したものはXの期待値である.

\begin{align}
\left.\frac{d}{ds}G_X(s)\right|_{s=1}&= \left.\frac{d}{ds}\mathbb{E}_{X\sim f_X}[s^X]\right|_{s=1}\\
&= \left.\frac{d}{ds}\sum_xs^xf_X(x)\right|_{s=1}\\
&= \left.\sum_xxs^{x-1}f_X(x)\right|_{s=1}\\
&= \sum_xxf_X(x)\\
&= \mathbb{E}_{X\sim f_X}[X]
\end{align}

同様に,これをk階微分し,s=1を代入したものはXのk次階乗モーメントである.

\begin{align}
\left.\frac{d^k}{ds^k}G_X(s)\right|_{s=1}&= \left.\frac{d^k}{ds^k}\mathbb{E}_{X\sim f_X}[s^X]\right|_{s=1}\\
&= \left.\frac{d^k}{ds^k}\sum_xs^xf_X(x)\right|_{s=1}\\
&= \left.\sum_x x(x-1)...(x-k+1)s^{x-k}f_X(x)\right|_{s=1}\\
&= \left.\sum_x x(x-1)...(x-k+1)s^{x-k}f_X(x)\right|_{s=1}\\
&= \sum_x x(x-1)...(x-k+1)f_X(x)\\
&= \mathbb{E}_{X\sim f_X}[X(X-1)...(X-k+1)]
\end{align}

答案

任意のx=0,1,...に対し,f_X(x)=>0が成り立つ.また,

\sum_{x=0}^\infty f_X(x)=\frac{1/2}{1-1/2}=1

より,fは確率関数の条件を満たす.確率関数fに従う確率変数Xの積率母関数は,

\begin{align}
M_X(t)&= \mathbb{E}_{X\sim f_X}[e^{tX}]\\
&= \sum_{x=0}^\infty \frac{e^{tx}}{2^{x+1}}\\
&= \sum_{x=0}^\infty\frac{1}{2}(\frac{e^t}{2})^x\\
&= \frac{1/2}{1-e^t/2}\\
&= \frac{1}{2-e^t}
\end{align}

また,Xの確率母関数は,

\begin{align}
G_X(s)&= \mathbb{E}_{X\sim f_X}[s^X]\\
&= \sum_{x=0}^\infty \frac{s^x}{2^{x+1}}\\
&= \sum_{x=0}^\infty\frac{1}{2}(\frac{s}{2})^x\\
&= \frac{1/2}{1-s/2}\\
&= \frac{1}{2-s}
\end{align}

確率母関数G_Xより,Xのk次階乗モーメントを求める.

\begin{align}
\frac{d}{ds}G_X(s)&= \frac{1}{(2-s)^2}\\
\frac{d^2}{ds^2}G_X(s)&= \frac{2(2-s)}{(2-s)^4}=\frac{2}{(2-s)^3}\\
\frac{d^3}{ds^3}G_X(s)&= \frac{3(2-s)^2\times 2}{(2-s)^6}=\frac{3!}{(2-s)^4}\\
\frac{d^4}{ds^4}G_X(s)&= \frac{4(2-s)^3\times 3!}{(2-s)^8}=\frac{4!}{(2-s)^5}\\
&\vdots\\
\frac{d^k}{ds^k}G_X(s)&= \frac{k!}{(2-s)^{k+1}}
\end{align}

従って,

\mathbb{E}[X(X-1)...(X-k+1)]=\left.\frac{d^k}{ds^k}G_X(s)\right|_{s=1}=k!

参考文献

2
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
2
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?