On the Identifiability of Finite Mixtures

1968/02/01 by Sidney J. Yakowitz, Sidney Yakowitz, John D. Spragins +1 · 399 citations
Computer Science · Mathematics · #Applied mathematics #Bayesian Methods and Mixture Models #Binomial (polynomial) #Cauchy distribution #Combinatorics #Exponential family #Finite field #Finite set #Gaussian #Identifiability #Mathematical analysis #Mathematics #Natural exponential family #Statistics

paper · pdf · doi:10.1214/aoms/1177698520

published in The Annals of Mathematical Statistics 39(1), 209-214 (Institute of Mathematical Statistics)

openalex publication_date 1968/02/01 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/26

Abstract

H. Teicher [5] has initiated a valuable study of the identifiability of finite mixtures (these terms to be defined in the next section), revealing a sufficiency condition that a class of finite mixtures be identifiable and from this, establishing the identifiability of all finite mixtures of one-dimensional Gaussian distributions and all finite mixtures of gamma distributions. From other considerations, he has generalized [4] a result of Feller [1] that arbitrary (and hence finite) mixtures of Poisson distributions are identifiable, and has also shown binomial and uniform families do not generate identifiable mixtures. In this paper it is proven that a family \mathscrF of cumulative distribution functions (cdf's) induces identifiable finite mixtures if and only if \mathscrF is linearly independent in its span over the field of real numbers. Also we demonstrate that finite mixtures of \mathscrF are identifiable if \mathscrF is any of the following: the family of n products of exponential distributions, the multivariate Gaussian family, the union of the last two families, the family of one-dimensional Cauchy distributions, and the non-degenerate members of the family of one-dimensional negative binomial distributions. Finally it is shown that the translation-parameter family generated by any one-dimensional cdf yields identifiable finite mixtures.

Cited by

• Uncovering Latent Reasoning Strategies in Language Models
• Reasoning Fine-Tuning Induces Persistent Latent Policy States
• Diffusion models recover accurate mixture weights despite score function insensitivity
• Nonparametric semi-supervised learning of class proportions
• Beyond ICA: Identifiability by Symmetry Breaking
• Identifiability of multivariate logistic mixture models
• Semiparametric Robust Estimation of Population Location
• Source Identification for Mixtures of Product Distributions
• From Stochasticity to Signal: A Bayesian Latent State Model for Reliable Measurement with LLMs
• Foundation Model Forecasts: Form and Function
• Strong consistency of pseudo-likelihood parameter estimator for univariate Gaussian mixture models
• Directed Cyclic Graph for Causal Discovery from Multivariate Functional Data
• Learning latent causal graphs via mixture oracles
• PMODE: Theoretically Grounded and Modular Mixture Modeling
• Haussdorff consistency of MLE in folded normal and Gaussian mixtures
• Learning Smooth Populations of Parameters with Trial Heterogeneity
• Diagonally-Weighted Generalized Method of Moments Estimation for Gaussian Mixture Modeling
• The Kernel-SME Filter for Multiple Target Tracking
• On consistency of the MLE under finite mixtures of location-scale distributions with a structural parameter
• Flow based approach for Dynamic Temporal Causal models with non-Gaussian or Heteroscedastic Noises
• Identifiability and consistency of network inference using the hub model and variants
• Finite mixture models do not reliably learn the number of components
• Neural Mixture Distributional Regression
• A Robust Monotonic Single-Index Model for Skewed and Heavy-Tailed Data: A Deep Neural Network Approach Applied to Periodontal Studies
• P-values for classification
• Mean Estimation from Coarse Data: Characterizations and Efficient Algorithms
• On the Robustness of Mixture Models in the Presence of Hidden Markov Regimes with Covariate-Dependent Transition Probabilities
• Inference for mixtures of symmetric distributions
• A Unifying Framework for Unsupervised Concept Extraction
• On the existence of fully inseparable biseparable Gaussian states
• Prediction Markets as Bayesian Inverse Problems: Uncertainty Quantification, Identifiability, and Information Gain from Price-Volume Histories under Latent Types
• Moment Varieties of Gaussian Mixtures
• Semiparametric robust mixture of experts based on nonparametric maximum likelihood
• Algebraic Identifiability of Gaussian Mixtures
• Parameters not empirically identifiable or distinguishable, including correlation between Gaussian observations
• Hybrid Probabilistic Zonotopes for Identifiable and Refinable Predictive Uncertainty
• Posterior Consistency in Parametric Models via a Tighter Notion of Identifiability
• The impact of covariance misspecification in multivariate Gaussian mixtures on estimation and inference: an application to longitudinal modeling. [europepmc]
• A COMPOSITIONAL MODEL TO ASSESS EXPRESSION CHANGES FROM SINGLE-CELL RNA-SEQ DATA. [europepmc]
• Mixture Complexity and Its Application to Gradual Clustering Change Detection. [europepmc]
• Directly selecting cell-type marker genes for single-cell clustering analyses. [europepmc]

Related

• On the Identifiability of Finite Mixtures of Finite Product Measures
• Generalized Identifiability Bounds for Mixture Models with Grouped Samples
• Identifiability and optimal rates of convergence for parameters of multiple types in finite mixtures
• On The Identifiability of Mixture Models from Grouped Samples
• A Criterion for Extending Continuous-Mixture Identifiability Results