【学习笔记】数理统计习题九

Q1: Let $X_1,\dots ,X_n$ be an iid sample of $N(\mu ,{\sigma }^2)$, where $\mu$ is known. Show that this model has a monotone likelihood ratio. Given a significance level $\alpha$, derive a UMP test of the following hypotheses:
$$H_0:{\sigma }^2\ge {\sigma }_0^{2}vs.H_1:{\sigma }^2<{\sigma }_0^2;$$$$H_0:{\sigma }^2\le {\sigma }_0^{2}vs.H_1:{\sigma }^2>{\sigma }_0^2.$$

解： 对于任意的${\sigma }_2^2>{\sigma }_1^2>0$，我们有 L ( x 1 : n ; σ 2 2 ) L ( x 1 : n ; σ 1 2 ) = ∏ i = 1 n ( 1 2 π σ 2 ) n e x p { − ( x i − μ ) 2 2 σ 2 2 } ( 1 2 π σ 1 ) n e x p { − ( x i − μ ) 2 2 σ 2 1 } = ( σ 1 σ 2 ) n e x p { 1 2 ( 1 σ 1 2 − 1 σ 2 2 ) ∑ i = 1 n ( x i − μ ) 2 } \frac{L(x_{1:n};\sigma_2^2)}{L(x_{1:n};\sigma_1^2)}=\displaystyle\prod_{i=1}^n\frac{(\frac{1}{\sqrt{2\pi}\sigma_2})^nexp\{-\frac{(x_i-\mu)^2}{2\sigma_2^2}\}}{(\frac{1}{\sqrt{2\pi}\sigma_1})^nexp\{-\frac{(x_i-\mu)^2}{2\sigma_2^1}\}}=(\frac{\sigma_{1}}{\sigma_2})^nexp\{\frac{1}{2}(\frac{1}{\sigma_1^2}-\frac{1}{\sigma_2^2})\displaystyle\sum_{i=1}^n(x_i-\mu)^2\} L(x1:n​;σ12​)L(x1:n​;σ22​)​=i=1∏n​(2π ​σ1​1​)nexp{ −2σ21​(xi​−μ)2​}(2π ​σ2​1​)nexp{ −2σ22​(xi​−μ)2​}​=(σ2​σ1​​)nexp{ 21​(σ12​1​−σ22​1​)i=1∑n​(xi​−μ)2}它可以表示为$T(x_{1:n})=\frac{1}{{\sigma }_0^2}{\sum }_{i=1}^n(x_i-\mu {)}^2$的递增函数。而且，当${\sigma }^2={\sigma }_0^2$时，$T(x_{1:n})=\frac{1}{{\sigma }_0^2}{\sum }_{i=1}^n(x_i-\mu {)}^2\sim {\chi }^2(n)$。因此，对于假设$$H_0:{\sigma }^2\ge {\sigma }_0^{2}vs.H_1:{\sigma }^2<{\sigma }_0^2;$$UMP拒绝域为 W I = { x 1 : n ∣ T ( x 1 : n ) < χ α 2 ( n ) } = { x 1 :