Approximate likelihood inference in generalized linear latent variable\n models based on integral dimension reduction

2015/03/04 by Silvia Bianconcini, Bianconcini, Silvia, Silvia Cagnone +3
Computer Science · Mathematics · #FOS: Computer and information sciences #Gaussian Processes and Bayesian Inference #Methodology (stat.ME) #Statistical Methods and Bayesian Inference #Statistical Methods and Inference

paper · pdf · doi:10.48550/arxiv.1503.01249

openalex publication_date 2015/03/04 · openalex created_date 2021/02/01 · openalex updated_date 2026/07/28

Abstract

Latent variable models represent a useful tool for the analysis of complex\ndata when the constructs of interest are not observable. A problem related to\nthese models is that the integrals involved in the likelihood function cannot\nbe solved analytically. We propose a computational approach, referred to as\nDimension Reduction Method (DRM), that consists of a dimension reduction of the\nmultidimensional integral that makes the computation feasible in situations in\nwhich the quadrature based methods are not applicable. We discuss the\nadvantages of DRM compared with other existing approximation procedures in\nterms of both computational feasibility of the method and asymptotic properties\nof the resulting estimators.\n

Citations

• Maximum Likelihood for Generalized Linear Models with Nested Random Effects via High-Order, Multivariate Laplace Approximation
• A General Approach to Confirmatory Maximum Likelihood Factor Analysis

Related

• Gradient-based data and parameter dimension reduction for Bayesian models: an information theoretic perspective
• Towards One Model for Classical Dimensionality Reduction: A Probabilistic Perspective on UMAP and t-SNE
• A unified performance analysis of likelihood-informed subspace methods
• Consistent Estimation of Low-Dimensional Latent Structure in High-Dimensional Data
• A sequential reduction method for inference in generalized linear mixed\n models