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

Estimating missing data: an iterative regression approach.

B Holt1, R A Benfer

  • 1Department of Anthropology, University of Missouri-Columbia, Columbia, MO, 65211, USA. holtmb@hotmail.com

Journal of Human Evolution
|August 31, 2000
PubMed
Summary
This summary is machine-generated.

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

Low failure to attend rates and increased clinic capacity with Telehealth: A highly effective outpatient model that should continue beyond the COVID-19 pandemic.

Journal of gastroenterology and hepatology·2020
Same author

Multipurpose prevention technologies: maximising positive synergies.

BJOG : an international journal of obstetrics and gynaecology·2014
Same author

Is Ginkgo biloba (Ginkgoaceae) really an oviparous plant?

American journal of botany·2011
Same author

The role of occlusal stress and gingival infection in the formation of exostoses on mandible and maxilla from Neolithic China.

Homo : internationale Zeitschrift fur die vergleichende Forschung am Menschen·2002
Same author

Creating senior-friendly Web sites.

Issue brief (Center for Medicare Education)·2002
Same author

Bootstrap methods for sex determination from the os coxae using the ID3 algorithm.

Journal of forensic sciences·2001
Same journal

The Omo-Turkana Basin fossil record as a milestone toward a comprehensive database of the hominin fossil record online.

Journal of human evolution·2026
Same journal

An early Late Pleistocene ochre from the Luonan Basin, central China.

Journal of human evolution·2026
Same journal

A new large-bodied sivaladapid primate from the Eocene of Myanmar resolves phylogenetic conflict obscuring the anthropoid status of Amphipithecidae.

Journal of human evolution·2026
Same journal

Bone retouchers and end-scrapers side by side at Bushman Rock Shelter: Exploring a ca. 100,000-year-old technological imbrication in southern Africa.

Journal of human evolution·2026
Same journal

A versatile gluteus medius drove a seamless transition to bipedalism.

Journal of human evolution·2026
Same journal

Dental morphology and ancient human dispersals within and out of Africa.

Journal of human evolution·2026
See all related articles

This study introduces MISDAT, an iterative regression method for estimating missing data. MISDAT offers improved accuracy over traditional techniques for scientific datasets.

Area of Science:

  • Anthropology
  • Bioinformatics
  • Statistics

Background:

  • Missing data is a pervasive issue across scientific disciplines.
  • Existing imputation methods like mean insertion and linear regression have limitations.
  • These methods can be problematic due to inherent biases or the nature of missing data patterns.

Purpose of the Study:

  • To introduce a novel, generalized iterative regression method for missing data estimation.
  • To demonstrate superior performance compared to traditional methods.
  • To provide a robust solution applicable to singular matrices and small datasets.

Main Methods:

  • Developed an iterative regression model (MISDAT) utilizing a bootstrap method.
  • Applied the method to primate anthropometric data with 20% simulated missing values.

Related Experiment Videos

  • Tested on European Upper Paleolithic and Mesolithic human postcranial measurements.

    • Main Results:

    • MISDAT demonstrated significantly better estimates in the first iteration compared to other techniques in simulation tests.
    • MISDAT outperformed mean imputation and classical multiple regression on human skeletal data.
    • Performance is comparable to classical multiple regression when squared multiple correlation values are high (approx. > 0.8).

    Conclusions:

    • The MISDAT method offers a more general and accurate approach to handling missing data in scientific research.
    • Its iterative bootstrap strategy enhances the precision of estimates.
    • Applicable to diverse datasets, including those with singular matrices or limited sample sizes.