That is, if $T$ is UMVUE for $\theta$, does this mean that $T$ is a complete and sufficient statistic for $\theta$?
$\begingroup$
$\endgroup$
1
-
1$\begingroup$ I believe the sample median of a sample drawn from the Laplace distribution provides a counter example but it needs to take me more time to write down a rigorous proof. I will come back when I got time. $\endgroup$– ZhanxiongCommented Aug 16, 2017 at 5:28
Add a comment
|