Characterizations of non-normalized discrete probability distributions and their application in statistics

Abstract

From the distributional characterizations that lie at the heart of Stein's method we derive explicit formulae for the mass functions of discrete probability laws that identify those distributions. These identities are applied to develop tools for the solution of statistical problems. Our characterizations, and hence the applications built on them, do not require any knowledge about normalization constants of the probability laws. To demonstrate that our statistical methods are sound, we provide comparative simulation studies for the testing of fit to the Poisson distribution and for parameter estimation of the negative binomial family when both parameters are unknown. We also consider the problem of parameter estimation for discrete exponential-polynomial models which generally are non-normalized.

24 pages, 3 figures

References in corpus (7)

• Interpretation and Generalization of Score Matching
• Generalized Score Matching for Non-Negative Data
• Tests based on characterizations, and their efficiencies: a survey
• Stein's method for discrete Gibbs measures
• Minimum Stein Discrepancy Estimators
• Minimum $L^q$-distance estimators for non-normalized parametric models
• Imputation estimators for unnormalized models with missing data