An unknown prior density g(θ) has yielded (unobservable) Θ1, Θ2, …, ΘN, and each Θi produces an observation Xi from an exponential family. deconvolveR is an R package for estimating prior distribution g(θ) from the data using Empirical Bayes deconvolution.

Details and examples may be found in the paper by Narasimhan and Efron, 2020. A vignette with further examples is also provided.