I am posting this question as a solution checking.

Let $X_1,...,X_{30}$ be a random sample from the exponential distribution with unknown mean $\mu\in \{1,1/\delta\}$ (where $\delta>1)$. Consider the best test of $H_0:\mu=1$ v.s. $H_1:\mu=1/\delta$ with power given by $\beta=\int_0^{17}\dfrac {\delta^{30}}{\Gamma(30)}y^{29}e^{-\delta y} dy$. What is the critical region of the test ?

My work: By Neyman-Pearson lemma, the best test is given by $L(1;\mathbb x)/L(1/\delta;\mathbb x)\le k$ i.e. $e^{-\sum x_i}/e^{-\delta \sum x_i} \le c$ i.e. $e^{(\delta-1)\sum x_i}\le c $ i.e. $\sum x_i \le c_1$ for some constant $c_1$. Now given $P_{H_1}(\sum_{i=1}^{30} X_i \le c_1)=\int_0^{17}\dfrac {\delta^{30}}{\Gamma(30)}y^{29}e^{-\delta y} dy$. Now under $H_1$, $X_i \sim exp (\delta)$ , so $\sum_{i=1}^{30} X_i \sim Gamma(30,1/\delta)$, hence $c_1=17$.

Thus the critical region is $\sum_{i=1}^{30} X_i\le 17$.

Am I correct ?

Please let me know.

Thanks in advance

  • $\begingroup$ Duplicate of question you asked on our math site. Seems you have double-posted several problems in this way. $\endgroup$ – BruceET Jan 15 at 20:59

Your Answer

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Browse other questions tagged or ask your own question.