EUVIMEDEUVIMED entdecken
Open Access

Causal K-means clustering

Kwangho Kim, Jisu Kim, Edward H Kennedy

Journal of the Royal Statistical Society Series B: Statistical Methodology · 12.11.2026 · 88 (5) · 1704-1721

DOI: 10.1093/jrsssb/qkag068 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of the Royal Statistical Society Series B: Statistical Methodology, online erschienen am 27.04.2026. Verfasst von Kwangho Kim und 2 weiteren Autor:innen. Beteiligte Einrichtung: Department of Statistics, Korea University, 145 Anam-ro, Seongbuk-gu, Seoul 02841, Korea. Die Arbeit wurde bisher 1-mal zitiert (Crossref) und nennt 59 Literaturangaben. Der Volltext ist frei in PubMed Central verfügbar. Gefördert von: NLM NIH HHS, National Research Foundation, MSIT.

AbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Causal effects are often characterized at the population level, which can mask important heterogeneity across latent subgroups. Since the subgroup structure is unknown, identifying and evaluating subgroup specific effects is substantially more challenging than standard population level analysis. We address this problem by proposing Causal k-Means Clustering, a framework that uses k-means clustering ideas to recover unknown subgroup structure from individual level causal contrasts. The problem differs fundamentally from ordinary clustering because the objects to be clustered are unknown counterfactual functions. We first study a simple plug in estimator that is readily implemented with standard algorithms and establish its rate of convergence. We then develop a bias-corrected estimator using semiparametric efficiency theory and double machine learning and show that it attains fast root-n rates and asymptotic normality in large nonparametric models. The proposed methods are especially useful for modern outcome wide studies with multiple treatment levels, and the framework extends naturally to clustering based on more general pseudo-outcomes, including partially observed outcomes and other unknown functionals. We study finite sample performance in simulations and illustrate the method with an application to mobile supported self-management for chronic low back pain.

Schlagwörter

Stichwörter der Autor:innen

Causal inferenceHeterogeneous treatment effectObservational studiesPersonalizationSubgroup analysis

Autor:innen und Institutionen

  • Department of Statistics, Korea University, 145 Anam-ro, Seongbuk-gu, Seoul 02841, Korea.
  • Department of Statistics, Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul 08826, Korea.
  • Department of Statistics and Data Science, Carnegie Mellon University, 5000 Forbes Ave, Pittsburgh, PA 15213, USA.

Förderung

  • NLM NIH HHS
  • National Research Foundation
  • MSIT
  • Samsung Science and Technology Foundation
  • National Library of Medicine
  • NSF

Zitieren

APA

Kim, K., Kim, J., & Kennedy, E. H. (2026). Causal K-means clustering. Journal of the Royal Statistical Society Series B: Statistical Methodology, 88(5), 1704-1721. https://doi.org/10.1093/jrsssb/qkag068

Vancouver

Kim K, Kim J, Kennedy EH. Causal K-means clustering. Journal of the Royal Statistical Society Series B: Statistical Methodology. 2026;88(5):1704-1721. doi:10.1093/jrsssb/qkag068

BibTeX
@article{Kim2026,
  author = {Kim, Kwangho and Kim, Jisu and Kennedy, Edward H},
  title = {Causal K-means clustering},
  journal = {Journal of the Royal Statistical Society Series B: Statistical Methodology},
  year = {2026},
  volume = {88},
  number = {5},
  pages = {1704--1721},
  doi = {10.1093/jrsssb/qkag068}
}
Kostenlos starten →