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

Multivariate outlier detection and remediation in geochemical databases.

G C Lalor1, C Zhang

  • 1International Centre for Environmental and Nuclear Sciences, University of the West Indies, Kingston, Jamaica.

The Science of the Total Environment
|January 10, 2002
PubMed
Summary

This study classifies outliers into range, spatial, and relationship types. Neural network methods effectively identify and correct relationship outliers in rare earth element soil data, improving environmental database quality control.

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

Ruthenium-catalyzed oxidative cleavage of olefins to aldehydes.

The Journal of organic chemistry·2001
Same author

Pharmacokinetics of testosterone after percutaneous gel or buccal administration.

Fertility and sterility·2001
Same author

The leukocyte common antigen-related protein tyrosine phosphatase receptor regulates regenerative neurite outgrowth in vivo.

The Journal of neuroscience : the official journal of the Society for Neuroscience·2001
Same author

Calpain activity in the rat brain after transient forebrain ischemia.

Experimental neurology·2001
Same author

Sharpening low-energy, standard-model tests via correlation coefficients in neutron beta decay.

Physical review letters·2001
Same author

Creation of chimeric mutant axolotls: a model to study early embryonic heart development in Mexican axolotls.

Anatomy and embryology·2001

Area of Science:

  • Geochemistry
  • Environmental Science
  • Data Science

Background:

  • Outlier detection is crucial for environmental data quality.
  • Rare earth element (REE) concentrations in soil require robust analytical methods.
  • Understanding spatial and correlational patterns in geochemical data is essential.

Purpose of the Study:

  • To classify and detect different types of outliers in geochemical datasets.
  • To evaluate the effectiveness of multivariate methods for outlier identification.
  • To demonstrate the application of neural networks for outlier remediation in environmental data.

Main Methods:

  • Classification of outliers into range, spatial, and relationship types.
  • Application of Principal Component Analysis (PCA), Multiple Regression Analysis (MRA), and Autoassociation Neural Network (AutoNN).

Related Experiment Videos

  • Utilizing a backpropagation neural network for predicting outlier 'expected values' and subsequent remeasurement for validation.

    • Main Results:

    • PCA effectively detected high-value range outliers.
    • AutoNN and MRA were successful in identifying relationship outliers.
    • Neural network prediction of low Sm concentrations in outliers was validated by remeasurement.

    Conclusions:

    • Neural network methods offer an automated and effective approach for environmental database quality control.
    • Model-free and non-linear problem-solving capabilities make neural networks suitable for complex geochemical data.
    • Accurate outlier detection and remediation enhance the reliability of environmental datasets.