Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Maximum-penalized-likelihood estimation for independent and Markov-dependent mixture models.

B G Leroux1, M L Puterman

  • 1Health and Welfare Canada, Environmental Health Centre, Ottawa, Ontario.

Biometrics
|June 1, 1992
PubMed
Summary

This study introduces maximum-penalized-likelihood methods for mixture models, enhancing parameter estimation and component selection. These techniques were successfully applied to fetal lamb movement data, revealing physiological mechanisms.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

A Prospective Cohort Study of Bisphenol A Exposure from Dental Treatment.

Journal of dental research·2020
Same author

Site-Specific Neutrophil Migration and CXCL2 Expression in Periodontal Tissue.

Journal of dental research·2016
Same author

Critique of reanalysis of Casa Pia data on associations of porphyrins and glutathione-S-transferases with dental amalgam exposure.

Human & experimental toxicology·2014
Same author

Computerized dental injection fear treatment: a randomized clinical trial.

Journal of dental research·2013
Same author

Area-based socio-economic disadvantage and tuberculosis incidence.

The international journal of tuberculosis and lung disease : the official journal of the International Union against Tuberculosis and Lung Disease·2012
Same author

Fractal dimension evaluation of cone beam computed tomography in patients with bisphosphonate-associated osteonecrosis.

Dento maxillo facial radiology·2011

Area of Science:

  • Statistics
  • Biostatistics
  • Computational Statistics

Background:

  • Mixture models are essential for analyzing complex data distributions.
  • Determining the optimal number of components and estimating parameters are critical challenges.
  • Existing methods may lack efficiency or robustness in certain applications.

Purpose of the Study:

  • To implement and evaluate maximum-penalized-likelihood procedures for mixture models.
  • To develop algorithms for automatic starting value generation for the EM algorithm.
  • To apply these methods to real-world physiological data.

Main Methods:

  • Maximum-penalized-likelihood estimation for independent and Markov-dependent mixture models.
  • Development of algorithms for automatic EM algorithm starting value generation.

Related Experiment Videos

  • Computation of the information matrix for parameter estimation.
  • Application of Poisson mixture models to ultrasound-derived fetal lamb movement counts.

    • Main Results:

    • The proposed methods effectively determine the number of mixing components.
    • Accurate estimation of model parameters was achieved.
    • The EM algorithm initialization was automated and efficient.
    • Poisson mixture models provided plausible explanations for fetal movement patterns.

    Conclusions:

    • Maximum-penalized-likelihood procedures offer a robust framework for mixture model analysis.
    • Automated EM algorithm initialization improves computational efficiency.
    • The application to fetal movement data demonstrates the practical utility in biostatistics.
    • The study contributes to understanding physiological processes through statistical modeling.