統計検定 1級 2022年 統計数理 問4 解答 解説

スポンサーリンク

[1]

\(f(x) = \displaystyle \frac{d}{dx}F(x)\)

\(\displaystyle =\left \{ \begin{array}{} \frac{1}{\gamma}x^{-\frac{1}{\gamma}-1} & (x>1) \\ 0 & (x \leq 1)\end{array} \right.\)

\(f_{\gamma=1}(x) \)\(\displaystyle =\left \{ \begin{array}{} x^{-2} & (x>1) \\ 0 & (x \leq 1)\end{array} \right.\)

[2]

\(E[X]\)

\(E[X] \displaystyle = \int^\infty_1 x \frac{1}{\gamma}x^{-\frac{1}{\gamma}-1}dx\)

\(\displaystyle = \int^\infty_1 \frac{1}{\gamma}x^{-\frac{1}{\gamma}}dx\)

\(\displaystyle =\left \{ \begin{array}{} \frac{1}{\gamma}[\ln x]^\infty _ 1 & (\gamma = 1) \\ \frac{1}{\gamma-1}\left[x^{1-\frac{1}{\gamma}}\right]^\infty _ 1 & (\gamma \neq 1)\end{array} \right.\)

\(\displaystyle =\left \{ \begin{array}{} \infty & (\gamma = 1) \\ \frac{1}{\gamma-1}\left[x^{\frac{\gamma-1}{\gamma}}\right]^\infty _ 1 & (\gamma > 1)\\ \frac{1}{\gamma-1}\left[\frac{1}{x^{\frac{1-\gamma}{\gamma}}}\right]^\infty _ 1 & (0<\gamma < 1)\end{array} \right.\)

\(\displaystyle =\left \{ \begin{array}{} \infty & (\gamma = 1) \\ \infty & (\gamma > 1)\\ \frac{1}{1-\gamma} & (0<\gamma < 1)\end{array} \right.\)

\(\displaystyle =\left \{ \begin{array}{} \frac{1}{1-\gamma} & (0<\gamma < 1) \\ \infty & (\gamma \geq 1) \end{array} \right.\)

\(V[X]\)

\(E[X]\)が収束する\(0<\gamma < 1\)の範囲で考える。

\(E[X^2] \displaystyle = \int^\infty_1 x^2 \frac{1}{\gamma}x^{-\frac{1}{\gamma}-1}dx\)

\(\displaystyle = \int^\infty_1 \frac{1}{\gamma}x^{\frac{\gamma-1}{\gamma}}dx\)

\(\displaystyle =\left \{ \begin{array}{} \frac{1}{2}[\ln x]^\infty _ 1 & \left(\gamma = \frac{1}{2}\right) \\ \frac{1}{2\gamma-1}\left[x^{\frac{2\gamma-1}{\gamma}}\right]^\infty _ 1 & \left(\gamma \neq \frac{1}{2}\right)\end{array} \right.\)

\(\displaystyle =\left \{ \begin{array}{} \infty &\left(\gamma = \frac{1}{2}\right) \\ \frac{1}{2\gamma-1}\left[x^{\frac{2\gamma-1}{\gamma}}\right]^\infty _ 1 & \left(\frac{1}{2}<\gamma<1\right)\\ \frac{1}{2\gamma-1}\left[\frac{1}{x^{\frac{1-2\gamma}{\gamma}}}\right]^\infty _ 1 & \left(0<\gamma<\frac{1}{2}\right)\end{array} \right.\)

\(\displaystyle =\left \{ \begin{array}{} \infty & \left(\frac{1}{2} \leq \gamma<1\right)\\ \frac{1}{1-2\gamma} & \left(0<\gamma<\frac{1}{2}\right)\end{array} \right.\)

ここで、公式から

\(V[X] = E[X^2]-E[X]^2\)

\(\displaystyle = \frac{1}{1-2\gamma} - \left( \frac{1}{1-\gamma}\right)^2\)\(\displaystyle = \frac{\gamma^2}{(1-2\gamma)(1-\gamma)^2} \)

[3]

尤度\(L\)は

\(\displaystyle L = \prod_{i=1}^n f(x_i)\)\(\displaystyle = \frac{1}{\gamma^n}\prod_{i=1}^n x_i^{-\frac{1}{\gamma}-1}\)

\(\displaystyle \ln L = -n \ln \gamma +\left(-\frac{1}{\gamma}-1\right)\sum _{i=1}^n \ln x_i \)

\(\displaystyle \frac{ \partial \ln L}{\partial \gamma} = -\frac{n}{\gamma} +\frac{\sum _{i=1}^n \ln x_i}{\gamma ^2} \)\(=0\)として、

\(\displaystyle \hat{\gamma} = \frac{1}{n}\sum _{i=1}^n \ln x_i\)

[4]

\(t = \displaystyle \frac{1}{\gamma}\ln x\)として、

\(x = e^{\gamma t}\)

\(dx = \gamma e^{\gamma t}dt\)と、[1]で求めた\(f(x)\)から、

\(T\)の確率密度関数\(g(t)\)は、

\(\displaystyle g(t) = \frac{1}{\gamma}e^{-\gamma t-t}\times \gamma e^{\gamma t}\)

\(\displaystyle = e^{-t} \) (\(T\)は指数分布に従う)

[5]

\(E[\hat{\gamma}]\)

\(\displaystyle E[\hat{\gamma}] = E\left[\frac{1}{n}\sum _{i=1}^n \ln x_i\right]\)

\(\displaystyle = \frac{1}{n}\sum _{i=1}^nE\left[ \ln x_i\right]\)

\(\displaystyle = \frac{1}{n}\sum _{i=1}^nE\left[\gamma T\right]\)

\(\displaystyle =\gamma E\left[ T\right] = \gamma\) (\(T\)は期待値\(1\)の指数分布に従うため)

\(V[\hat{\gamma}]\)

\(\displaystyle V[\hat{\gamma}] = V\left[\frac{1}{n}\sum _{i=1}^n \ln x_i\right]\)

\(\displaystyle = \frac{1}{n^2}\sum _{i=1}^nV\left[ \ln x_i\right]\)

\(\displaystyle = \frac{1}{n^2}\sum _{i=1}^nV\left[\gamma T\right]\)

\(\displaystyle =\frac{\gamma^2}{n}V\left[ T\right] \)\(\displaystyle =\frac{\gamma^2}{n}\) (\(T\)は分散\(1\)の指数分布に従うため)


コメント

タイトルとURLをコピーしました