1
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.

3.9(標準) ポアソン分布

Posted at

方針

ポアソン分布について書く.ポアソン分布は単位時間内に平均lambda回起きる事象が起こる回数を示す確率分布である.確率密度関数は,

f_X(x)=\frac{\lambda^x}{x!}\exp(-\lambda),\ \ \ x=0,...

積率母関数は,

\begin{align}
M_X(t)&= \mathbb{E}_{X\sim Po(\lambda)}[e^{tX}]\\
&= \sum_{x=0}^\infty\frac{\lambda^x}{x!}\exp(tx-\lambda)\\
&= \exp(-\lambda)\sum_{x=0}^\infty \frac{(e^t\lambda)^x}{x!}\\
&= \exp(-\lambda)\exp(e^t\lambda)\\
&= \exp\{\lambda(e^t-1)\}
\end{align}

二項分布の積率母関数(3.8で解説)において,

\lambda=np

と置くと,n→∞において

\begin{align}
\lim_{n\rightarrow\infty}(\frac{\lambda}{n}e^t+1-\frac{\lambda}{n})^n&= \lim_{n\rightarrow\infty}\{1+\frac{\lambda(e^t-1)}{n}\}^n\\
&= \exp\{\lambda(e^t-1)\}
\end{align}

であるから,ポアソン分布は二項分布でn→∞とした形であることがわかる.

期待値は,

\begin{align}
\mathbb{E}_{X\sim Po(\lambda)}[X]&= \sum_{x=0}^\infty x\frac{\lambda^x}{x!}\exp(-\lambda)\\
&= \lambda\sum_{x=1}^\infty\frac{\lambda^{x-1}}{(x-1)!}\exp(-\lambda)\\
&= \lambda\sum_{x'=0}^\infty\frac{\lambda^{x'}}{x'!}\exp(-\lambda)\ \ \ \ x'=x-1\\
&= \lambda
\end{align}

分散は,

\begin{align}
\mathbb{E}_{X\sim Po(\lambda)}[X^2]&= \sum_{x=0}^\infty x^2\frac{\lambda^x}{x!}\exp(-\lambda)\\
&= \lambda\sum_{x=1}^\infty x\frac{\lambda^{x-1}}{(x-1)!}\exp(–\lambda)\\
&= \lambda\sum_{x'=0}^\infty(x'+1)\frac{\lambda^{x'}}{x'!}\exp(-\lambda)\\
&= \lambda(\lambda+1)\\
Var(X)&= \lambda(\lambda+1)-\lambda^2=\lambda
\end{align}

答案

確率変数Yの積率母関数は,

\begin{align}
M_Y(t)&= \mathbb{E}_{X\sim Po(\lambda)}[e^{tY}]\\
&= \mathbb{E}_{X\sim Po(\lambda)}[e^{\frac{t(X-\lambda)}{\sqrt{\lambda}}}]\\
&= \exp(-t\sqrt{\lambda})\mathbb{E}_{X\sim Po(\lambda)}[e^{\frac{t}{\sqrt{\lambda}}X}]\\
&= \exp(-t\sqrt{\lambda})\sum_{x=0}^\infty\frac{\lambda^x}{x!}\exp(\frac{t}{\sqrt{\lambda}}x-\lambda)\\
&= \exp(-t\sqrt{\lambda}-\lambda)\sum_{x=0}^\infty\frac{(e^{\frac{t}{\sqrt{\lambda}}})^x}{x!}\\
&= \exp(-t\sqrt{\lambda}-\lambda+\lambda e^{\frac{t}{\sqrt{\lambda}}})
\end{align}

今,

\psi(t)=\log M_Y(t)

とおけば,n→∞で,

\psi(t)\rightarrow\frac{t^2}{2}

であることを示せば良い.

\psi(t))=-t\sqrt{\lambda}-\lambda+\lambda e^{\frac{t}{\sqrt{\lambda}}}

指数部分について,漸近展開して,

e^{\frac{t}{\sqrt{\lambda}}}=1+\frac{t}{\sqrt{\lambda}}+\frac{t^2}{2\lambda}+\mathcal{o}(\frac{1}{\lambda})\ \ \ as\ \ \lambda\rightarrow\infty

なので,

\begin{align}
\psi(t)&= -t\sqrt{\lambda}-\lambda+\lambda(1+\frac{t}{\sqrt{\lambda}}+\frac{t^2}{2\lambda}+\mathcal{o}(\frac{1}{\lambda}))\\
&= -t\sqrt{\lambda}-\lambda+\lambda+t\sqrt{\lambda}+\frac{t^2}{2}+\mathcal{o}(1)\\
&\approx \frac{t^2}{2}\ \ \ as\ \lambda\rightarrow\infty
\end{align}

参考文献

1
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
1
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?