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.

2.12(標準) 積率母関数2, 変数変換3

Last updated at Posted at 2021-07-17

1.方針

積率母関数を定義にそって求める.一様分布の平均,分散については定義通り計算した方が早いが,練習のため積率母関数を経由して求める.

1.答案

Xの積率母関数は,

\begin{align}
M_X(t)&= \mathbb{E}_{X\sim f_X}[e^{tX}]\\
&= \int_0^1e^{tx}dx\\
&= \left[\frac{e^{tx}}{t}\right]_0^1\\
&= \frac{e^t-1}{t}
\end{align}

である.次に,これを一階微分すると,

\frac{d}{dt}M_X(t)=\frac{te^t-e^t+1}{t^2} 

である.従って,Xの平均は,

\begin{align}
\mathbb{E}_{X\sim f_X}[X]&= \lim_{t\rightarrow0}\frac{d}{dt}M_X(t)\\
&= \lim_{t\rightarrow 0}\frac{te^t-e^t+1}{t^2}\\
&= \lim_{t\rightarrow0}\frac{e^t+te^t-e^t}{2t}\\
&= \lim_{t\rightarrow0}\frac{e^t(t+1)}{2}=\frac{1}{2}
\end{align}

次に,積率母関数を二階微分すると,

\begin{align}
\frac{d^2}{dt^2}M_X(t)&= \frac{(-e^t+e^t+te^t)t^2-2t(-e^t+1+te^t)}{t^4}\\
&= \frac{t^2e^t-2(-e^t+1+te^t)}{t^3}\\
&= \frac{t^2e^t-2te^t+2e^t-2}{t^3}
\end{align}

なので,Xの二次モーメントは

\begin{align}
\mathbb{E}_{X\sim f_X}[X^2]&= \lim_{t\rightarrow0}\frac{d^2}{dt^2}M_X(t)\\
&= \lim_{t\rightarrow0}\frac{2te^t+t^2e^t-2e^t-2te^t+2e^t}{3t^2}\\
&= \lim_{t\rightarrow0}\frac{t^2e^t}{3t^2}\\
&= \lim_{t\rightarrow0}\frac{2te^t+t^2e^t}{6t}\\
&= \lim_{t\rightarrow0}\frac{2e^t+2te^t+2te^t+t^2e^t}{6}=\frac{1}{3}
\end{align}

以上より,Xの分散は

Var(X)=\mathbb{E}_{X\sim f_X}[X^2]-(\mathbb{E}_{X\sim f_X}[X])^2=\frac{1}{12}

2.方針

変数変換はまず分布関数から考える.分布関数ができたらそれを微分することで確率密度関数を求める.

2.答案

Yの分布関数は,

\begin{align}
F_Y(y)&= P[Y\leq y]\\
&= P[X^2\leq y]\\
&= P[X\leq\sqrt{y}]\\
&= \int_0^{\sqrt{y}}1sx\\
&= \left[x\right]_0^{\sqrt{y}}=\sqrt{y}\ \ \ (0<y<1)
\end{align}

なので,確率密度関数は

f_Y(y)=\frac{d}{dy}F_Y(y)=\frac{1}{2\sqrt{y}}

また,平均と分散は,

\begin{align}
\mathbb{E}_{Y\sim f_Y}[Y]&= \int_0^1\frac{y}{2\sqrt{y}}dy\\
&= \int_0^1\frac{y^{1/2}}{2}dy\\
&= \left[\frac{1}{2}\frac{2}{3}y^{3/2}\right]_0^1=\frac{1}{3}\\
\mathbb{E}_{Y\sim f_Y}[Y^2]&= \int_0^1\frac{y^2}{2\sqrt{y}}dy\\
&= \int_0^1\frac{y^{3/2}}{2}dy\\
&= \left[\frac{1}{2}\frac{2}{5}y^{5/2}\right]_0^1=\frac{1}{5}\\
Var(Y)&= \mathbb{E}_{Y\sim f_Y}[Y^2]-(\mathbb{E}_{Y\sim f_Y}[Y])^2\\
&= \frac{1}{5}-\frac{1}{9}=\frac{4}{45}
\end{align}

3.方針

変数変換はまず分布関数から考える.分布関数ができたらそれを微分することで確率密度関数を求める.

3.答案

Yの分布関数は,

\begin{align}
F_Y(y)&= P[Y\leq y]\\
&= P[-\log X\leq y]\\
&= P[\log X\geq -y]\\
&= P[X\geq e^{-y}]\\
&= \int_{e^{-y}}^1f_X(x)dx\\
&= \left[x\right]_{e^{-y}}^1=1-e^{-y}\ \ (0<y)
\end{align}

なので,確率密度関数は,

f_Y(y)=e^{-y}

また,平均と分散は

\begin{align}
\mathbb{E}_{Y\sim f_Y}[Y]&= \int_0^\infty y\exp(-y)dy\\
&= \int_0^\infty y\{-\exp(-y)\}'dy\\
&= \left[-y\exp(-y)\right]_0^\infty+\int_0^\infty\exp(-y)dy\\
&= 1\\
\mathbb{E}_{Y\sim f_Y}[Y^2]&= \int_0^\infty y^2\exp(-y)dy\\
&= \int_0^\infty y^2\{-\exp(-y)\}'dy\\
&= \left[-y^2\exp(-y)\right]_0^\infty+2\int_0^\infty y\exp(-y)dy\\
&= 2\\
Var(Y)&= \mathbb{E}_{Y\sim f_Y}[Y^2]-(\mathbb{E}_{Y\sim f_Y}[Y])^2\\
&= 2-1=1
\end{align}

4.方針

変数変換はまず分布関数から考える.分布関数ができたらそれを微分することで確率密度関数を求める.

4.答案

Yの分布関数は

\begin{align}
F_Y(y)&= P[Y\leq y]\\
&= P[\sigma X+\mu\leq y]\\
&= P[X\leq \frac{y-\mu}{\sigma}]\ \ (\because\sigma>0)\\
&= \int_0^{\frac{y-\mu}{\sigma}}dx\\
&= \left[x\right]_0^{\frac{y-\mu}{\sigma}}=\frac{y-\mu}{\sigma}\ \ \ (\mu<y<\sigma+\mu)
\end{align}

なので,確率密度関数は,

f_Y(y)=\frac{d}{dy}F_Y(y)=\frac{1}{\sigma}

積率母関数は,

\begin{align}
M_Y(t)&= \mathbb{E}_{Y\sim f_Y}[e^{tY}]\\
&= \int_\mu^{\sigma+\mu}\frac{e^{ty}}{\sigma}dy\\
&= \left[\frac{e^{ty}}{\sigma t}\right]_\mu^{\sigma+\mu}\\
&= \frac{1}{\sigma t}(e^{t(\sigma+\mu)}-e^{t\mu})
\end{align}

平均と分散は,

\begin{align}
\mathbb{E}_{Y\sim f_Y}[Y]&= \int_\mu^{\sigma+\mu}\frac{y}{\sigma}dy\\
&= \left[\frac{1}{2\sigma}y^2\right]_\mu^{\sigma+\mu}= \frac{\sigma}{2}+\mu\\
\mathbb{E}_{Y\sim f_Y}[Y^2]&= \int_\mu^{\sigma+\mu}\frac{y^2}{\sigma}dy\\
&= \left[\frac{1}{3\sigma}y^3\right]_\mu^{\sigma+\mu}=\frac{\sigma^2}{3}+\mu^2+\mu\sigma\\
Var(Y)&= \mathbb{E}_{Y\sim f_Y}[Y^2]-(\mathbb{E}_{Y\sim f_Y}[Y])^2\\
&= \frac{\sigma^2}{3}+\mu^2+\mu\sigma-\frac{\sigma^2}{4}-\mu\sigma-\mu^2=\frac{\sigma^2}{12}
\end{align}

参考文献

1
0
3

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?