Decompounding discrete distributions

28 Mar 2019 / Reading time: 1 min

Check here for a new preprint on Bayesian decompounding for discrete distributions written by Ester Mariucci, Frank van der Meulen and me. The paper is a nice mixture of theory and practice. On the practical side, a computational algorithm for sampling from the posterior distribution is proposed and its good performance on simulated and real datasets is demonstrated (the Julia code is available here). On the theoretical side, posterior consistency of the method is established and the corresponding posterior contraction rate is derived. Working on this paper was fun: how often do you get an opportunity in statistics to solve Diophantine equations?