New exponential dispersion models for count data: the ABM and LM classes

DOI10.1051/PS/2021001zbMATH Open1481.60018arXiv2003.13849OpenAlexW3120520217WikidataQ115554261 ScholiaQ115554261MaRDI QIDQ4990908FDOQ4990908

Authors: Shaul K. Bar-Lev, A. Ridder 

Publication date: 1 June 2021

Published in: ESAIM: Probability and Statistics (Search for Journal in Brave)

Abstract: In their fundamental paper on cubic variance functions, Letac and Mora (The Annals of Statistics,1990) presented a systematic, rigorous and comprehensive study of natural exponential families on the real line, their characterization through their variance functions and mean value parameterization. They presented a section that for some reason has been left unnoticed. This section deals with the construction of variance functions associated with natural exponential families of counting distributions on the set of nonnegative integers and allows to find the corresponding generating measures. As exponential dispersion models are based on natural exponential families, we introduce in this paper two new classes of exponential dispersion models based on their results. For these classes, which are associated with simple variance functions, we derive their mean value parameterization and their associated generating measures. We also prove that they have some desirable properties. Both classes are shown to be overdispersed and zero-inflated in ascending order, making them as competitive statistical models for those in use in both, statistical and actuarial modeling. To our best knowledge, the classes of counting distributions we present in this paper, have not been introduced or discussed before in the literature. To show that our classes can serve as competitive statistical models for those in use (e.g., Poisson, Negative binomial), we include a numerical example of real data. In this example, we compare the performance of our classes with relevant competitive models.

Full work available at URL: https://arxiv.org/abs/2003.13849

Recommendations

• Exponential dispersion models for overdispersed zero-inflated count data
• scientific article; zbMATH DE number 7762724
• On Poisson-exponential-Tweedie models for ultra-overdispersed count data
• A new infinitely divisible discrete distribution with applications to count data modeling
• An empirical model for underdispersed count data
• The binomial-discrete Poisson-Lindley model: modeling and applications to count regression
• Mean-parametrized Conway–Maxwell–Poisson regression models for dispersed counts
• On \(q\)-exponential count models
• A New Discrete Distribution: Properties, Characterizations, Modeling Real Count Data, Bayesian and Non-Bayesian Estimations

zbMATH Keywords

overdispersionnatural exponential familyvariance functionexponential dispersion modelzero-inflated distribution

Mathematics Subject Classification ID

Probability distributions: general theory (60E05) Computational methods for problems pertaining to statistics (62-08) Characterization and structure theory of statistical distributions (62E10)

Cites Work

• Title not available (Why is that?)
• On Hinde-Demétrio regression models for overdispersed count data
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Overdispersion: Models and estimation.
• Natural exponential families with quadratic variance functions
• Title not available (Why is that?)
• A suitable discrete distribution for modelling automobile claim frequencies
• On Strict Arcsine Distribution
• Title not available (Why is that?)
• The discrete Lindley distribution: properties and applications
• Discrete dispersion models and their Tweedie asymptotics
• A new discrete distribution with actuarial applications
• A cohort-based extension to the Lee-Carter model for mortality reduction factors
• Some discrete exponencial dispersion models: Poisson-Tweedie and Hinde-Demétrio classes
• Natural real exponential families with cubic variance functions
• Reproducibility and natural exponential families with power variance functions
• A Property of Count Distributions in the Hinde–Demétrio Family
• Extended Poisson–Tweedie: Properties and regression models for count data
• On the mean value parametrization of natural exponential families -- a revisited review
• Modelling claim number using a new mixture model: negative binomial gamma distribution
• Geometric Tweedie regression models for continuous and semicontinuous data with variation phenomenon
• On the two-parameter Bell–Touchard discrete distribution
• A new infinitely divisible discrete distribution with applications to count data modeling

Cited In (4)

• A delineation of new classes of exponential dispersion models supported on the set of nonnegative integers
• Title not available (Why is that?)
• Exponential dispersion models for overdispersed zero-inflated count data
• Some discrete exponencial dispersion models: Poisson-Tweedie and Hinde-Demétrio classes

This page was built for publication: New exponential dispersion models for count data: the ABM and LM classes

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4990908)