Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
October 1, 2001Biometrika

Testing the number of components in a normal mixture

View Full Paper
Ask AI
Bookmark
Share

Authors

YLYungtai LoAlbert Einstein College of Medicine

Discussion

Loading...

Member takes

Overview

Methodological study demonstrates an adjusted likelihood ratio test for normal mixtures, indicating reliable determination of latent subpopulation counts.

Key Points

  • To establish a formal statistical hypothesis test for determining the exact number of distinct components in a normal mixture distribution.
  • Derived the asymptotic distribution of the standard likelihood ratio test statistic when comparing k versus k+1 normal mixture components.
  • Evaluated the proposed test statistic against nonstandard regularity condition constraints inherent in mixture modeling.
  • Provides an adjusted likelihood ratio test distribution specifically tailored for finite normal mixtures where regular asymptotic theory fails.
  • Demonstrates reliable type I error control and sufficient statistical power to distinguish true subpopulation structures in continuous data.

Cite This Study

Yungtai Lo (2001) studied this question.

synapsesocial.com/papers/696aa0fb72d15f53efa1e660https://doi.org/10.1093/biomet/88.3.767
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Likelihood Ratio Tests in Contamination Models1999 · 47 citations
  2. 2Note on the Consistency of the Maximum Likelihood Estimate for Nonidentifiable Distributions1981 · 197 citations
  3. 3Likelihood Ratio Tests for Model Selection and Non-Nested Hypotheses1989 · 6,071 citations
  4. 4Maximum Likelihood from Incomplete Data Via the <i>EM</i> Algorithm1977 · 49,948 citations
  5. 5Asymptotic Properties of Maximum Likelihood Estimators and Likelihood Ratio Tests under Nonstandard Conditions1987 · 2,575 citations