Computation for latent variable model estimation: a unified stochastic proximal framework

DOI10.1007/S11336-022-09863-9zbMATH Open1499.62441arXiv2008.07214OpenAlexW3049206795MaRDI QIDQ2103576FDOQ2103576

Siliang Zhang, Yunxiao Chen

Publication date: 9 December 2022

Published in: Psychometrika (Search for Journal in Brave)

Abstract: Latent variable models have been playing a central role in psychometrics and related fields. In many modern applications, the inference based on latent variable models involves one or several of the following features: (1) the presence of many latent variables, (2) the observed and latent variables being continuous, discrete, or a combination of both, (3) constraints on parameters, and (4) penalties on parameters to impose model parsimony. The estimation often involves maximizing an objective function based on a marginal likelihood/pseudo-likelihood, possibly with constraints and/or penalties on parameters. Solving this optimization problem is highly non-trivial, due to the complexities brought by the features mentioned above. Although several efficient algorithms have been proposed, there lacks a unified computational framework that takes all these features into account. In this paper, we fill the gap. Specifically, we provide a unified formulation for the optimization problem and then propose a quasi-Newton stochastic proximal algorithm. Theoretical properties of the proposed algorithms are established. The computational efficiency and robustness are shown by simulation studies under various settings for latent variable model estimation.

Full work available at URL: https://arxiv.org/abs/2008.07214

Recommendations

• Simulation-based approach to estimation of latent variable models
• Latent Variable Modelling: A Survey*
• scientific article; zbMATH DE number 1264868
• ALMOND: Adaptive Latent Modeling and Optimization via Neural Networks and Langevin Diffusion
• Asymptotic properties of adaptive maximum likelihood estimators in latent variable models

zbMATH Keywords

latent variable modelsstochastic approximationproximal algorithmpenalized estimatorquasi-Newton methodsPolyak-Ruppert averaging

Mathematics Subject Classification ID

Point estimation (62F10) Factor analysis and principal components; correspondence analysis (62H25) Applications of statistics to psychology (62P15) Stochastic approximation (62L20)

Cites Work

• A penalty approach to differential item functioning in Rasch models
• The elements of statistical learning. Data mining, inference, and prediction
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Statistics for high-dimensional data. Methods, theory and applications.
• A numerically stable dual method for solving strictly convex quadratic programs
• On the convergence properties of the EM algorithm
• Latent Variable Models and Factor Analysis
• Title not available (Why is that?)
• Variational Analysis
• Regularization and Variable Selection Via the Elastic Net
• Measurement Error in Nonlinear Models
• Consistent Estimates Based on Partially Consistent Observations
• Maximum likelihood estimates in exponential response models
• Introductory lectures on convex optimization. A basic course.
• Title not available (Why is that?)
• Title not available (Why is that?)
• Title not available (Why is that?)
• Acceleration of Stochastic Approximation by Averaging
• A simple general method for oblique rotation
• Adaptive Rejection Sampling for Gibbs Sampling
• Title not available (Why is that?)
• A Stochastic Approximation Method
• Title not available (Why is that?)
• Stability of Stochastic Approximation under Verifiable Conditions
• On the ergodicity properties of some adaptive MCMC algorithms
• Analysis of Panel Data
• Introduction to Stochastic Search and Optimization
• A composite likelihood inference in latent variable models for ordinal longitudinal responses
• MCMC estimation and some model-fit analysis of multidimensional IRT models
• Title not available (Why is that?)
• Multidimensional item reponse theory.
• Compound decision theory and empirical Bayes methods
• Title not available (Why is that?)
• Title not available (Why is that?)
• On Asymptotic Normality in Stochastic Approximation
• On perturbed proximal gradient algorithms
• Title not available (Why is that?)
• Dynamic Latent Trait Models for Multidimensional Longitudinal Data
• Statistical Analysis ofQ-Matrix Based Diagnostic Classification Models
• The generalized DINA model framework
• A Markov chain Monte Carlo approach to confirmatory item factor analysis
• High-dimensional exploratory item factor analysis by a Metropolis-Hastings Robbins-Monro algorithm
• The stochastic EM algorithm: Estimation and asymptotic results
• Stochastic Gradient Descent on Riemannian Manifolds
• Stochastic Methods for Composite and Weakly Convex Optimization Problems
• Convergence of adaptive and interacting Markov chain Monte Carlo algorithms
• Proximal Newton-Type Methods for Minimizing Composite Functions
• A stochastic approximation algorithm with Markov chain Monte-Carlo method for incomplete data estimation problems
• Fitting nonlinear ordinary differential equation models with random effects and unknown initial conditions using the stochastic approximation expectation-maximization (SAEM) algorithm
• Inconsistent maximum likelihood estimators for the Rasch model
• Latent variable selection for multidimensional item response theory models via \(L_{1}\) regularization
• Identifiability of restricted latent class models with binary responses
• Joint maximum likelihood estimation for high-dimensional exploratory item factor analysis
• An improved stochastic <scp>EM</scp> algorithm for large‐scale full‐information item factor analysis
• On single versus multiple imputation for a class of stochastic algorithms estimating maximum likelihood
• Convergence of Markovian stochastic approximation with discontinuous dynamics
• Structured Latent Factor Analysis for Large-scale Data: Identifiability, Estimability, and Their Implications
• A note on exploratory item factor analysis by singular value decomposition
• Robust measurement via a fused latent and graphical item response theory model

Cited In (6)

• Scalable Bayesian approach for the DINA Q-matrix estimation combining stochastic optimization and variational inference
• ALMOND: Adaptive Latent Modeling and Optimization via Neural Networks and Langevin Diffusion
• A generalized expectation model selection algorithm for latent variable selection in multidimensional item response theory models
• Fast and universal estimation of latent variable models using extended variational approximations
• Fast estimation of generalized linear latent variable models for performance and process data with ordinal, continuous, and count observed variables
• Stochastic proximal-gradient algorithms for penalized mixed models

Uses Software

• quadprog
• DIFlasso

This page was built for publication: Computation for latent variable model estimation: a unified stochastic proximal framework

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2103576)