---
title: "Dempster-Laird-Rubin EM algorithm"
type: "entity"
entity_kind: "concept"
status: "hub"
canonical_name: "Dempster-Laird-Rubin EM algorithm"
aliases: ["EM algorithm (Dempster, Laird & Rubin 1977)","Dempster, Laird & Rubin (1977)","expectation-maximization"]
first_seen: "2026-08-14T00:00:00.000Z"
connects_to: ["Dawid-Skene model","maximum likelihood estimation","latent-class models","A. Philip Dawid","incomplete-data estimation"]
writer_model: "claude-sonnet-5"
seek_code_commit: "17d9798"
---


A general iterative method for maximum-likelihood estimation from incomplete data, published in 1977 by Arthur Dempster, Nan Laird, and Donald Rubin ("Maximum likelihood from incomplete data via the EM algorithm," *Journal of the Royal Statistical Society, Series B*, 39, 1-38) — one of the most widely cited papers in statistics, later applied across latent-variable modeling, clustering, and machine learning generally. Matters to this vault as the numerical method [[entity-dawid-skene-model|Dawid & Skene (1979)]] explicitly credits and depends on to fit their own observer-error-rate model: the previously-unread ancestor standing one citation hop behind a lineage the vault had already traced from a 1979 clinical-medicine paper through 2014 crowdsourcing research to 2025 LLM retrieval.

## References
- 2026-08-14: Dawid & Skene (1979) name this paper directly as supplying "a numerical method of maximum likelihood estimation which is ideally suited to" their own problem — extending the vault's Dawid-Skene citation lineage one hop further back. ([[claim-dawid-skene-1979-credits-dempster-laird-rubin-1977-for-em-method]])