題組內容
3. (30%) Let X₁, X₂, ..., Xₙ be a random sample from Bernoulli(p). Define R(p, δ(X)) = Eₚ[(δ(X) - p)²] (the Mean-Square Error).
(d) Consider three estimators
$$\hat{p}_U = \text{UMVUE found in (b)},$$
$\hat{p}_B$ = the posterior mean with constant risk found in (c-iii),
and
$$\hat{p}T = \begin{cases} \hat{p}U \text{ with probability } \frac{n}{n+1}, \\ 1/2 \text{ with probability } \frac{1}{n+1}. \end{cases}$$
(d-iv) \( (3\%) \) Show that \( \hat{p}_U \) is asymptotically efficient of \( p \).