A method to handle zero counts in the multinomial model

Title: A method to handle zero counts in the multinomial model
Creator: Tuyl, Frank
Relation: The American Statistician Vol. 73, Issue 2, p. 151-158
Publisher Link: http://dx.doi.org/10.1080/00031305.2018.1444673
Publisher: American Statistical Association
Resource Type: journal article
Date: 2019
Description: In the context of an objective Bayesian approach to the multinomial model, Dirichlet(a, …, a) priors with a < 1 have previously been shown to be inadequate in the presence of zero counts, suggesting that the uniform prior (a = 1) is the preferred candidate. In the presence of many zero counts, however, this prior may not be satisfactory either. A model selection approach is proposed, allowing for the possibility of zero parameters corresponding to zero count categories. This approach results in a posterior mixture of Dirichlet distributions and marginal mixtures of beta distributions, which seem to avoid the problems that potentially result from the various proposed Dirichlet priors, in particular in the context of extreme data with zero counts.
Subject: Bayesian inference; Bayes-Laplace prior; model and variable selection; objective priors; spike and slab prior
Identifier: http://hdl.handle.net/1959.13/1451064
Identifier: uon:44084
Identifier: ISSN:0003-1305
Language: eng
Reviewed

		Thumbnail	File	Description	Size	Format