所屬科目:研究所、轉學考(插大)◆數理統計學
(a) For every fixed \( \varepsilon > 1 \), find a distribution for \( X \) with \( E(X) = 0 \) that gives equality in Markov's inequality. (10 points)
(b) Prove for an arbitrary random variable \( X \) $$P(|X| \geq \varepsilon) \leq \frac{E(\cosh(X)) - 1}{\cosh(\varepsilon) - 1}.$$ Here $\cosh(x) = (e^x + e^{-x})/2$. (10 points)
2. Let $X_1, \ldots, X_n$ be a sample from a population with the density $$f(x) = \theta(\theta + 1)x^{\theta-1}(1 - x),$$ $0 < x < 1, \theta > 0$. Find a method of moments estimate of $\theta$. (10 points)
3. Let \( \mathbf{X} = (X_1, \ldots, X_n) \) be a sample from a \( U(0, \theta) \) population, where \( 0 < \theta < \infty \) and \( \theta \) is unknown. Then find the U.M.V.U. estimate of \( \theta \). (15 points)
4. Suppose that \( X_{1}, \ldots, X_{n} \) is a sample from the pdf, \( f(x|\theta) = \exp(-(x - \theta)) \), \( x > \theta \) and \( \theta \in \Re \). Then construct the U.M.P. level \( \alpha \) test for $$H_0 : \theta \leq \theta_0 \text{ vs. } H_1 : \theta > \theta_0.$$ (15 points)
(a) Show \( [\bar{X} - \sigma z(1 - \alpha_1) / \sqrt{n}, \bar{X} + \sigma z(1 - \alpha_2) / \sqrt{n}] \) is a level \( (1 - \alpha) \) confidence interval for \( \mu \), where \( \bar{X} = \sum_{i=1}^{n} X_i / n \) and \( \alpha_1 + \alpha_2 = \alpha \). (10 points)
(b) From (a), the shortest level \( (1 - \alpha) \) interval is obtained by taking \( \alpha_{1} = \alpha_{2} = \alpha / 2 \). (10 points)
(a) Find the maximum likelihood estimate of \( p, \hat{p} \). (10 pints)
(b) Find the asymptotic distribution of \( \sqrt{n} (\hat{p} - p) \). (10 points)