Wavelet shrinkage for natural exponential families with quadratic variance functions

We propose a wavelet shrinkage methodology for univariate natural exponential families with quadratic variance functions, covering the Gaussian, Poisson, gamma, binomial, negative binomial and generalised hyperbolic secant distributions. Simulation studies for Poisson and binomial data are used to illustrate the usefulness of the proposed methodology, and comparisons are made with other methods available in the literature. We also present applications to datasets arising from high-energy astrophysics and from epidemiology.