Mixtures of power series distributions: identifiability via uniqueness in problems of moments

We treat the identifiability problem for mixtures involving power series distributions. Applying an idea of Sapatinas (Ann Inst Stat Math 47:447–459, 1995) we prove and elaborate that a mixture distribution is identifiable if a certain Stieltjes problem of moments has a unique solution while a non-uniqueness leads to a non- identifiable mixture. We describe explicitly models of identifiable mixtures and models of non-identifiable mixtures. Illustrative examples and comments on related questions are also given.

[1]  O. Barndorff-Nielsen.  Identifiability of mixtures of exponential families , 1965 .

[2]  William Feller,et al.  An Introduction to Probability Theory and Its Applications , 1967 .

[3]  The identifiability of mixtures of distributions , 1969 .

[4]  B. Everitt,et al.  Finite Mixture Distributions , 1981 .

[5]  A. F. Smith,et al.  Statistical analysis of finite mixture distributions , 1986 .

[6]  U. Lüxmann-Ellinghaus,et al.  On the identifiability of mixtures of infinitely divisible power series distributions , 1987 .

[7]  M. Meerschaert.  Regular Variation in R k , 1988 .

[8]  Jordan Stoyanov,et al.  Counterexamples in Probability , 1988 .

[9]  Yu. A. Brychkov,et al.  Integrals and series , 1992 .

[10]  Eric V. Slud,et al.  The Moment Problem for Polynomial Forms in Normal Random Variables , 1993 .

[11]  T. Rao,et al.  Identifiability in stochastic models : characterization of probability distributions , 1994 .

[12]  Jacek Wesolowski,et al.  Bivariate distributions via a Pareto conditional distribution and a regression function , 1995 .

[13]  J. Wesolowski,et al.  Bivariate discrete measures via a power series conditional distribution and a regression function , 1995 .

[14]  B. Lindsay.  Mixture models : theory, geometry, and applications , 1995 .

[15]  T. Sapatinas.  Identifiability of mixtures of power-series distributions and related characterizations , 1995, Annals of the Institute of Statistical Mathematics.

[16]  J. Wesolowski,et al.  Posterior mean identifies the prior distribution in nb and related models , 1997 .

[17]  Gwo Dong Lin,et al.  On the moment problems , 1997 .

[18]  Jian-Lun Xu.  Characterizations of the exponential distribution via the blocking time in a queueing system with an unreliable server , 1998 .

[19]  Jordan Stoyanov,et al.  Krein condition in probabilistic moment problems , 2000 .

[20]  Geoffrey J. McLachlan,et al.  Finite Mixture Models , 2019, Annual Review of Statistics and Its Application.

[21]  Jong-Wuu Wu,et al.  Criteria for the Unique Determination of Probability Distributions by Moments , 2001 .

[22]  A. Pakes,et al.  Remarks on converse Carleman and Krein criteria for the classical moment problem , 2001, Journal of the Australian Mathematical Society.

[23]  Arjun K. Gupta,et al.  Identifiability of Modified Power Series Mixtures via Posterior Means , 2001 .

[24]  Allan Gut,et al.  The Moment Problem , 2002, Encyclopedia of Special Functions: The Askey-Bateman Project.

[25]  On the moment determinacy of the distributions of compound geometric sums , 2002, Journal of Applied Probability.

[26]  Hajo Holzmann,et al.  Identifiability of finite mixtures - with applications to circular distributions , 2004 .

[27]  Jordan Stoyanov,et al.  Stieltjes classes for moment-indeterminate probability distributions , 2004, Journal of Applied Probability.

[28]  Jordan Stoyanov,et al.  Method for constructing Stieltjes classes for M-indeterminate probability distributions , 2005, Appl. Math. Comput..

[29]  Dankmar Böhning,et al.  Asymptotic Normality in Mixtures of Power Series Distributions , 2005 .