Hypothesis test for normal mixture models: The EM approach

Abstract

Normal mixture distributions are arguably the most important mixture models, and also the most technically challenging. The likelihood function of the normal mixture model is unbounded based on a set of random samples, unless an artificial bound is placed on its component variance parameter. Moreover, the model is not strongly identifiable so it is hard to differentiate between over dispersion caused by the presence of a mixture and that caused by a large variance, and it has infinite Fisher information with respect to mixing proportions. There has been extensive research on finite normal mixture models, but much of it addresses merely consistency of the point estimation or useful practical procedures, and many results require undesirable restrictions on the parameter space. We show that an EM-test for homogeneity is effective at overcoming many challenges in the context of finite normal mixtures. We find that the limiting distribution of the EM-test is a simple function of the $0.5{\chi }_0^2+0.5{\chi }_1^2$ and ${\chi }_1^2$ distributions when the mixing variances are equal but unknown and the ${\chi }_2^2$ when variances are unequal and unknown. Simulations show that the limiting distributions approximate the finite sample distribution satisfactorily. Two genetic examples are used to illustrate the application of the EM-test.

Published in at http://dx.doi.org/10.1214/08-AOS651 the Annals of Statistics (http://www.imstat.org/aos/) by the Institute of Mathematical Statistics (http://www.imstat.org)

References in corpus (1)

• Estimation and confidence sets for sparse normal mixtures

Cited by in corpus (29)

• Universal Inference
• Galactic Double Neutron Star total masses and Gaussian mixture model selection
• Hypothesis testing with e-values
• Singularity, Misspecification, and the Convergence Rate of EM
• Another Argument in Favour of Wilcoxon's Signed Rank Test
• A Scalable Empirical Bayes Approach to Variable Selection in Generalized Linear Models
• Influential Feature PCA for high dimensional clustering
• Hypothesis testing for quantitative trait locus effects in both location and scale in genetic backcross studies
• Sharp Analysis of Expectation-Maximization for Weakly Identifiable Models
• On the Minimax Optimality of the EM Algorithm for Learning Two-Component Mixed Linear Regression
• Non-asymptotic detection of two-component mixtures with unknown means
• Semiparametric Mixtures of Regressions with Single-index for Model Based Clustering
• Gaussian Mixture Clustering Using Relative Tests of Fit
• Estimating a mixing distribution on the sphere using predictive recursion
• Penalized Maximum Likelihood Estimator for Skew Normal Mixtures
• Variance-Reduced Stochastic Optimization for Efficient Inference of Hidden Markov Models
• Lower Bounds on the Total Variation Distance Between Mixtures of Two Gaussians
• Homogeneity testing under finite location-scale mixtures
• Testing Homogeneity for Normal Mixture Models: Variational Bayes Approach
• An Overview of Semiparametric Extensions of Finite Mixture Models
• A mixture distribution approach for assessing genetic impact from twin study
• Detecting SNPs with interactive effects on a quantitative trait
• Testing Homogeneity in a heteroscedastic contaminated normal mixture
• Addressing patient heterogeneity in disease predictive model development
• Asymptotic Analysis of the Bayesian Likelihood Ratio for Testing Homogeneity in Normal Mixture Models
• On the Variance of the Fisher Information for Deep Learning
• Hypothesis test on a mixture forward-incubation-time epidemic model with application to COVID-19 outbreak
• Latent Gaussian Mixture Models for Nationwide Kidney Transplant Center Evaluation
• Test for the statistical significance of a treatment effect in the presence of hidden sub-populations