EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-ND 4.0

Spectral Clustering with Likelihood Refinement for High-dimensional Latent Class Recovery

Zhongyuan Lyu, Yuqi Gu iD

Psychometrika · 2026

DOI: 10.1017/psy.2026.10095 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Psychometrika, 2026. Verfasst von Zhongyuan Lyu und 1 weiteren Autor:innen. Beteiligte Einrichtung: https://ror.org/00hj8s172Columbia University, USA. Hauptthemen laut MEDLINE: Latent Class Analysis. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-ND 4.0). Gefördert von: Division of Mathematical Sciences.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Latent class models (LCMs) are widely used for identifying unobserved subgroups from multivariate categorical data in social sciences, with binary data as a particularly popular example. However, accurately recovering individual latent class memberships remains challenging, especially when handling high-dimensional datasets with many items. This work proposes a novel two-stage algorithm for LCMs suited for high-dimensional binary responses. Our method first initializes latent class assignments by an easy-to-implement spectral clustering algorithm, and then refines these assignments with a one-step likelihood-based update. This approach combines the computational efficiency of spectral clustering with the improved statistical accuracy of likelihood-based estimation. We establish theoretical guarantees showing that this method is minimax-optimal for latent class recovery in the statistical decision theory sense. The method also leads to exact clustering of subjects with high probability under mild conditions. As a byproduct, we propose a computationally efficient consistent estimator for the number of latent classes. Extensive experiments on both simulated data and real data validate our theoretical results and demonstrate our method's superior performance over alternative methods.

Schlagwörter

MeSH-Schlagwörter (MEDLINE)

Latent Class AnalysisLikelihood FunctionsClustering AlgorithmsAlgorithmsCluster AnalysisHumansComputer SimulationModels, Statistical

Stichwörter der Autor:innen

high-dimensional datajoint maximum likelihood estimationlatent class modelspectral clustering

Autor:innen und Institutionen

  • https://ror.org/00hj8s172Columbia University, USA.
  • Department of Statistics, https://ror.org/00hj8s172Columbia University, USA.

ORCID-Profile: Yuqi Gu

Förderung

  • Division of Mathematical Sciences

Zitieren

APA

Lyu, Z., & Gu, Y. (2026). Spectral Clustering with Likelihood Refinement for High-dimensional Latent Class Recovery. Psychometrika. https://doi.org/10.1017/psy.2026.10095

Vancouver

Lyu Z, Gu Y. Spectral Clustering with Likelihood Refinement for High-dimensional Latent Class Recovery. Psychometrika. 2026. doi:10.1017/psy.2026.10095

BibTeX
@article{Lyu2026,
  author = {Lyu, Zhongyuan and Gu, Yuqi},
  title = {Spectral Clustering with Likelihood Refinement for High-dimensional Latent Class Recovery},
  journal = {Psychometrika},
  year = {2026},
  doi = {10.1017/psy.2026.10095}
}
Kostenlos starten →