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

Methods of selecting informative variables.

Valerii V Fedorov1, Agnes M Herzberg, Sergei L Leonov

  • 1GlaxoSmithKline, Collegeville, PA 19426-0989, USA. Valeri.V.Fedorov@gsk.com

Biometrical Journal. Biometrische Zeitschrift
|March 21, 2006
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

Comparisons of minimization and Atkinson's algorithm.

Statistics in medicine·2010
Same author

Two-stage design for dose-finding that accounts for both efficacy and safety.

Statistics in medicine·2008
Same author

Dose finding designs for continuous responses and binary utility.

Journal of biopharmaceutical statistics·2007
Same author

Modelling, prediction and adaptive adjustment of recruitment in multicentre trials.

Statistics in medicine·2007
Same author

Practical usage of O'Brien's OLS and GLS statistics in clinical trials.

Pharmaceutical statistics·2007
Same author

Optimal design for dose response using beta distributed responses.

Journal of biopharmaceutical statistics·2005
Same journal

Comparison of Different Methods for the Meta-Analysis of Diagnostic Test Accuracy Studies-A Simulation Study.

Biometrical journal. Biometrische Zeitschrift·2026
Same journal

When to Adjust for Multiple Testing: A Unifying Guiding Principle.

Biometrical journal. Biometrische Zeitschrift·2026
Same journal

Ensuring Quality in Preclinical Research: The Importance of Being Human.

Biometrical journal. Biometrische Zeitschrift·2026
Same journal

Addressing Cluster-Level Treatment Effect Heterogeneity in Sample Size Determination for Hierarchical 2 × 2 Factorial Designs.

Biometrical journal. Biometrische Zeitschrift·2026
Same journal

A Multiple Imputation Approach to Distinguish Curative From Life-Prolonging Effects in the Presence of Missing Covariates.

Biometrical journal. Biometrische Zeitschrift·2026
Same journal

Tests for Categorical Data Beyond Pearson: A Distance Covariance and Energy Distance Approach.

Biometrical journal. Biometrische Zeitschrift·2026
See all related articles

This study introduces a novel method for selecting the most informative variables, incorporating measurement costs and accounting for variability and errors. The approach offers scale-invariant variable selection, outperforming existing dimension reduction techniques.

Area of Science:

  • Statistics
  • Data Science
  • Biostatistics

Background:

  • Variable selection is crucial for building accurate predictive models.
  • Existing dimension reduction methods like Principal Component Analysis (PCA) can be sensitive to scale transformations.
  • Incorporating measurement costs and accounting for inherent data variability are practical challenges in variable selection.

Purpose of the Study:

  • To develop a new, robust method for selecting the most informative variables from directly measurable sets.
  • To create a variable selection approach that is invariant to scale transformations.
  • To integrate measurement costs into the variable selection process for enhanced practicality.

Main Methods:

  • Utilizing information metrics analogous to experimental design theory, such as the determinant of the prediction dispersion matrix.

Related Experiment Videos

  • Developing algorithms inspired by optimal experimental design principles to handle population variability and observational errors.
  • Implementing a scale-invariant variable selection technique.

    • Main Results:

    • The proposed method demonstrates superior performance compared to traditional dimension reduction techniques (e.g., PCA, principal variables, battery reduction) on clinical data.
    • The variable selection process is shown to be invariant to scale transformations, a key advantage over other methods.
    • The approach effectively incorporates measurement costs, enhancing its practical applicability.

    Conclusions:

    • The novel variable selection method offers a robust and practical solution for identifying informative variables.
    • Its scale-invariant nature and consideration of measurement costs provide significant advantages over existing approaches.
    • This method holds promise for applications in various fields, particularly in analyzing complex datasets like clinical data.