Standard

A note on a computationally efficient implementation of the EM algorithm in item response models. / Robitzsch, Alexander.

In: Quantitative and Computational Methods in Behavioral Sciences, Vol. 1, No. 1, e3783, 11.05.2021.

Research output: Contribution to journalJournal articleResearchpeer-review

Harvard

Robitzsch, A 2021, 'A note on a computationally efficient implementation of the EM algorithm in item response models', Quantitative and Computational Methods in Behavioral Sciences, vol. 1, no. 1, e3783. https://doi.org/10.5964/qcmb.3783

APA

Robitzsch, A. (2021). A note on a computationally efficient implementation of the EM algorithm in item response models. Quantitative and Computational Methods in Behavioral Sciences, 1(1), [e3783]. https://doi.org/10.5964/qcmb.3783

Vancouver

Robitzsch A. A note on a computationally efficient implementation of the EM algorithm in item response models. Quantitative and Computational Methods in Behavioral Sciences. 2021 May 11;1(1). e3783. https://doi.org/10.5964/qcmb.3783

Author

Robitzsch, Alexander. / A note on a computationally efficient implementation of the EM algorithm in item response models. In: Quantitative and Computational Methods in Behavioral Sciences. 2021 ; Vol. 1, No. 1.

BibTeX

@article{40dae735b1de4051b96202f27575e2d2,
title = "A note on a computationally efficient implementation of the EM algorithm in item response models",
abstract = "This note sketches two computational shortcuts for estimating unidimensional item response models and multidimensional item response models with between-item dimensionality utilizing an expectation-maximization (EM) algorithm that relies on numerical integration with fixed quadrature points. It is shown that the number of operations required in the E-step can be reduced in situations of many cases and many items by appropriate shortcuts. Consequently, software implementations of a modified E-step in the EM algorithm could benefit from gains in computation time.",
keywords = "Methodological research and method development, computation time, statistical software, item response model, EM algorithm",
author = "Alexander Robitzsch",
year = "2021",
month = may,
day = "11",
doi = "https://doi.org/10.5964/qcmb.3783",
language = "English",
volume = "1",
journal = "Quantitative and Computational Methods in Behavioral Sciences",
issn = "2699-8432",
publisher = "PsychOpen",
number = "1",

}

RIS

TY - JOUR

T1 - A note on a computationally efficient implementation of the EM algorithm in item response models

AU - Robitzsch, Alexander

PY - 2021/5/11

Y1 - 2021/5/11

N2 - This note sketches two computational shortcuts for estimating unidimensional item response models and multidimensional item response models with between-item dimensionality utilizing an expectation-maximization (EM) algorithm that relies on numerical integration with fixed quadrature points. It is shown that the number of operations required in the E-step can be reduced in situations of many cases and many items by appropriate shortcuts. Consequently, software implementations of a modified E-step in the EM algorithm could benefit from gains in computation time.

AB - This note sketches two computational shortcuts for estimating unidimensional item response models and multidimensional item response models with between-item dimensionality utilizing an expectation-maximization (EM) algorithm that relies on numerical integration with fixed quadrature points. It is shown that the number of operations required in the E-step can be reduced in situations of many cases and many items by appropriate shortcuts. Consequently, software implementations of a modified E-step in the EM algorithm could benefit from gains in computation time.

KW - Methodological research and method development

KW - computation time

KW - statistical software

KW - item response model

KW - EM algorithm

U2 - https://doi.org/10.5964/qcmb.3783

DO - https://doi.org/10.5964/qcmb.3783

M3 - Journal article

VL - 1

JO - Quantitative and Computational Methods in Behavioral Sciences

JF - Quantitative and Computational Methods in Behavioral Sciences

SN - 2699-8432

IS - 1

M1 - e3783

ER -

ID: 1593842