Empirical Bayes methods for learning prior distributions from data. An unknown prior distribution (g) has yielded (unobservable) parameters, each of which produces a data point from a parametric exponential family (f). The goal is to estimate the unknown prior ("g-modeling") by deconvolution and Empirical Bayes methods.
An unknown prior density $g(\theta)$ has yielded (unobservable) $\Theta_1, \Theta_2,\ldots,\Theta_N$, and each $\Theta_i$ produces
an observation $X_i$ from an exponential family.
deconvolveR is an R package for estimating prior distribution $g(\theta)$ from the data
using Empirical Bayes deconvolution.
The current package is still under construction but will soon appear on CRAN along with a manuscript. Meanwhile, you can reproduce many examples by installing the package in R thus:
devtools::install_github("bnaras/deconvolveR") library(deconvolveR) vignette("deconvolution")
ignoreZero = TRUEby default makes the thinning correction. Please note this change which is also discussed in the vignette.