REMLA: An R package for robust expectation-maximization estimation for latent variable models

Factor analysis is a widely used statistical method for describing a large number of observed, correlated variables in terms of a smaller number of unobserved variables. Applications of this method usually impose the same latent variable model on all individuals in the sample, but this assumption mi...

Full description

Saved in:
Bibliographic Details
Main Authors: Kenneth J. Nieser, Bryan Saúl Ortiz-Torres, Gabriel Zayas-Cabán, Amy Cochran
Format: Article
Language:English
Published: Elsevier 2025-05-01
Series:SoftwareX
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2352711025000792
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849326156725092352
author Kenneth J. Nieser
Bryan Saúl Ortiz-Torres
Gabriel Zayas-Cabán
Amy Cochran
author_facet Kenneth J. Nieser
Bryan Saúl Ortiz-Torres
Gabriel Zayas-Cabán
Amy Cochran
author_sort Kenneth J. Nieser
collection DOAJ
description Factor analysis is a widely used statistical method for describing a large number of observed, correlated variables in terms of a smaller number of unobserved variables. Applications of this method usually impose the same latent variable model on all individuals in the sample, but this assumption might not hold as individuals can differ in attributes (e.g., age, gender) that influence model parameters. REMLA is an R package that implements a robust expectation–maximization (REM) algorithm to estimate the parameters for factor analysis models in a way that automatically acknowledges, and even detects, differences among individuals within the sample. This paper explains the methodological background of the estimation process, describes the algorithms employed, and illustrates how REMLA can be used to perform exploratory and confirmatory factor analyses through examples. In the future, we plan to extend this package to other latent variable models, such as mixture models.
format Article
id doaj-art-93c8f435226a407eabe9ebd2609823fc
institution Kabale University
issn 2352-7110
language English
publishDate 2025-05-01
publisher Elsevier
record_format Article
series SoftwareX
spelling doaj-art-93c8f435226a407eabe9ebd2609823fc2025-08-20T03:48:13ZengElsevierSoftwareX2352-71102025-05-013010211210.1016/j.softx.2025.102112REMLA: An R package for robust expectation-maximization estimation for latent variable modelsKenneth J. Nieser0Bryan Saúl Ortiz-Torres1Gabriel Zayas-Cabán2Amy Cochran3Center for Innovation to Implementation, VA Palo Alto Health Care System, Menlo Park, CA, United States of America; Stanford-Surgery Policy Improvement Research and Education Center, Department of Surgery, Stanford University, Stanford, CA, United States of America; Correspondence to: 795 Willow Road, Bldg. 324, Menlo Park, CA, United States of America.Department of Industrial and Systems Engineering, University of Wisconsin–Madison, Madison, WI, United States of AmericaDepartment of Industrial and Systems Engineering, University of Wisconsin–Madison, Madison, WI, United States of AmericaDepartment of Population Health Sciences, University of Wisconsin–Madison, Madison, WI, United States of America; Department of Mathematics, University of Wisconsin–Madison, Madison, WI, United States of AmericaFactor analysis is a widely used statistical method for describing a large number of observed, correlated variables in terms of a smaller number of unobserved variables. Applications of this method usually impose the same latent variable model on all individuals in the sample, but this assumption might not hold as individuals can differ in attributes (e.g., age, gender) that influence model parameters. REMLA is an R package that implements a robust expectation–maximization (REM) algorithm to estimate the parameters for factor analysis models in a way that automatically acknowledges, and even detects, differences among individuals within the sample. This paper explains the methodological background of the estimation process, describes the algorithms employed, and illustrates how REMLA can be used to perform exploratory and confirmatory factor analyses through examples. In the future, we plan to extend this package to other latent variable models, such as mixture models.http://www.sciencedirect.com/science/article/pii/S2352711025000792Expectation–maximization algorithmRobust statisticsStructural equation modelingLatent variablesFactor analysis
spellingShingle Kenneth J. Nieser
Bryan Saúl Ortiz-Torres
Gabriel Zayas-Cabán
Amy Cochran
REMLA: An R package for robust expectation-maximization estimation for latent variable models
SoftwareX
Expectation–maximization algorithm
Robust statistics
Structural equation modeling
Latent variables
Factor analysis
title REMLA: An R package for robust expectation-maximization estimation for latent variable models
title_full REMLA: An R package for robust expectation-maximization estimation for latent variable models
title_fullStr REMLA: An R package for robust expectation-maximization estimation for latent variable models
title_full_unstemmed REMLA: An R package for robust expectation-maximization estimation for latent variable models
title_short REMLA: An R package for robust expectation-maximization estimation for latent variable models
title_sort remla an r package for robust expectation maximization estimation for latent variable models
topic Expectation–maximization algorithm
Robust statistics
Structural equation modeling
Latent variables
Factor analysis
url http://www.sciencedirect.com/science/article/pii/S2352711025000792
work_keys_str_mv AT kennethjnieser remlaanrpackageforrobustexpectationmaximizationestimationforlatentvariablemodels
AT bryansaulortiztorres remlaanrpackageforrobustexpectationmaximizationestimationforlatentvariablemodels
AT gabrielzayascaban remlaanrpackageforrobustexpectationmaximizationestimationforlatentvariablemodels
AT amycochran remlaanrpackageforrobustexpectationmaximizationestimationforlatentvariablemodels