Accelerating L 1 -penalized expectation maximization algorithm for latent variable selection in multidimensional two-parameter logistic models

One of the main concerns in multidimensional item response theory (MIRT) is to detect the relationship between observed items and latent traits, which is typically addressed by the exploratory analysis and factor rotation techniques. Recently, an EM-based L 1 -penalized log-likelihood method (EML1)...

Full description

Saved in:
Bibliographic Details
Published inPloS one Vol. 18; no. 1
Main Authors Shang, Laixu, Ping-Feng, Xu, Na Shan, Man-Lai, Tang, Ho, George
Format Journal Article
LanguageEnglish
Published Public Library of Science 17.01.2023
Subjects
Algorithms
Computer applications
Feature selection
Generalized linear models
Maximum likelihood method
Optimization
Quantitative psychology
Rotation
Simulation
Online AccessGet full text

Cover

More Information
Summary:One of the main concerns in multidimensional item response theory (MIRT) is to detect the relationship between observed items and latent traits, which is typically addressed by the exploratory analysis and factor rotation techniques. Recently, an EM-based L 1 -penalized log-likelihood method (EML1) is proposed as a vital alternative to factor rotation. Based on the observed test response data, EML1 can yield a sparse and interpretable estimate of the loading matrix. However, EML1 suffers from high computational burden. In this paper, we consider the coordinate descent algorithm to optimize a new weighted log-likelihood, and consequently propose an improved EML1 (IEML1) which is more than 30 times faster than EML1. The performance of IEML1 is evaluated through simulation studies and an application on a real data set related to the Eysenck Personality Questionnaire is used to demonstrate our methodologies.
ISSN:1932-6203
DOI:10.1371/journal.pone.0279918