Exact non-Gaussian SURE bias and transformed shrinkage criteria
Statement
Two approaches to generalize Stein's method to non-Gaussian distributions are the following: \begin{itemize} \item[(i)] Stein coefficients: for a random variable , the \emph{Stein coefficient} satisfies [\mathbb{E}(Y f(Y)) = \mathbb{E}(T f'(Y)).] \item[(ii)] Zero-biasing: for of mean 0 and variance , compute the that satisfies [\mathbb{E}(Yf(Y)) = \sigma^2 \mathbb{E} f'(Y^*).]
Let , and let . Suppose that , and let . Here we will assume is a known distribution while is unknown.
Consider estimators of of the form . Recall that SURE for Gaussian random variables gives the unbiased estimate of risk [\hat{r}_h = d + 2\nabla \cdot h(X) + |h(X)|^2.] Calculate the bias [e_d := \mathbb{E}(\hat{r}_h - \mathbb{E} |\hat{\theta} - \theta|^2)] using Stein coefficients or zero-biasing. Does converge to 0 as ? Can we apply this to wavelet shrinkage? \end{itemize}
Record
- Source
- Added
Comments
Nothing has been published against this problem yet, and nothing has been said about it.
Sign in with an institutional address to take part in the discussion. Reading every thread stays open to everyone.
Sign inSolve with an agent
Open the statement in a chat, with the problem and the ground rules already written into the prompt.