Abstract:Mixture models are widely used to explain excessive variation in observations that is not captured by standard parametric models, and they lead to suggestive latent structures. The hypothetical latent structure often needs critical examination based on experimental data. It is therefore important to know the sample size needed to ensure a reasonable chance of success. We investigate this issue for the EM-test and the test. They are shown to be asymptotically equivalent and have simple limiting distributions under two sets of local alternatives for commonly used mixture models. We obtain a simple sample-size formula and an associated simulation-based calibration procedure, and we demonstrate via data examples and simulation studies that they provide useful guidance for several common mixture models. The Canadian Journal of Statistics 44: 82-101; 2016 (c) 2016 Statistical Society of CanadaResumeLes modeles de melange sont largement utilises pour expliquer des variations que les modeles parametriques habituels n'arrivent pas a capturer, et ils suggerent ainsi une structure latente qui doit souvent faire l'objet d'un examen critique base sur des donnees. Il importe donc de connai tre la taille d'echantillon necessaire afin de garantir une probabilite de succes raisonnable pour un tel examen. Les auteurs etudient ce probleme pour les tests EM et C(). Ils montrent que ces deux tests sont asymptotiquement equivalents et qu'ils presentent une distribution limite simple sous des hypotheses locales pour des modeles de melange frequemment utilises. Les auteurs obtiennent une formule simple pour la taille d'echantillon pourvue d'une procedure de calibration basee sur la simulation. Ils demontrent a l'aide d'exemples de donnees et de simulations que leur approche s'avere utile pour de nombreux modeles de melange communs. La revue canadienne de statistique 44: 82-101; 2016 (c) 2016 Societe statistique du Canada

Estimation and Hypothesis-Testing in Finite Mixture-Models

Estimation and hypothesis testing in finite mixture models

Bayesian estimation and prediction for certain mixtures

Hypothesis Testing in Mixture Regression Models

Hypothesis Testing in Mixture Regression Models ( Mathematical Details )

Testing the Order of Multivariate Normal Mixture Models

Homogeneity testing under finite location-scale mixtures

Mixture Models With a Prior on the Number of Components

A New Approach to Parameter Estimation of Mixture of Two Normal Distributions

Sample‐size Calculation for Tests of Homogeneity

Statistical inference for a mixture of Markov jump processes

Semiparametric Inference in a Genetic Mixture Model

On the Behavior of the Expectation-Maximization Algorithm for Mixture Models

Maximum-penalized-likelihood estimation for independent and Markov-dependent mixture models.

Testing Homogeneity in A Multivariate Mixture Model

Nonparametric Mixture of Regression Models.

M-Testing Using Finite and Infinite Dimensional Parameter Estimators

A smoothed semiparametric likelihood for estimation of nonparametric finite mixture models with a copula-based dependence structure

Asymptotics for Markov chain mixture detection

Full-model estimation for non-parametric multivariate finite mixture models

Statistical Inference on Component Parameters in Mixture Location-Scale Distribution