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

Classifiers as a model-free group comparison test.

Bommae Kim1, Timo von Oertzen2

  • 1Federal Reserve Bank of Kansas City, Kansas City, USA. bommae.kim@kc.frb.org.

Behavior Research Methods
|April 5, 2017
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

Automated emotion recognition via video-based semantic embeddings.

Frontiers in digital health·2026
Same author

Problematic pornography use and psychological distress: A longitudinal study in a large US sample.

Addictive behaviors·2025
Same author

Corrigendum to "Child maltreatment in young adults with residential youth care background: Prevalence and post-placement trends" [Child Abuse Negl. 157C (2024) 107074].

Child abuse & neglect·2025
Same author

Child maltreatment in young adults with residential youth care background: Prevalence and post-placement trends.

Child abuse & neglect·2024
Same author

Implementing machine learning techniques for continuous emotion prediction from uniformly segmented voice recordings.

Frontiers in psychology·2024
Same author

Products of Variables in Structural Equation Models.

Structural equation modeling : a multidisciplinary journal·2023
Same journal

Exploring psychological tradeoffs: Developing and demonstrating an R Shiny app for Pareto optimization.

Behavior research methods·2026
Same journal

The performance of Bayesian fit measures in detecting misspecified multilevel structural equation modeling.

Behavior research methods·2026
Same journal

Psychometric functions from multiple responses : Dedicated to the memory of Colin L. Mallows.

Behavior research methods·2026
Same journal

Low-cost, open-source, full-stack software and Arduino-based hardware for control of commercially available animal behavior systems.

Behavior research methods·2026
Same journal

PyNeon: A Python package for the analysis of Neon multimodal mobile eye-tracking data.

Behavior research methods·2026
Same journal

Talking surveys: How photorealistic embodied conversational agents shape response quality, engagement, and satisfaction.

Behavior research methods·2026
See all related articles

This study introduces machine learning classification algorithms for group comparison, offering a model-free approach. Support vector machines (SVMs) detect group differences effectively, especially with multiple variables, complementing traditional statistical tests.

Area of Science:

  • Statistics
  • Machine Learning
  • Data Science

Background:

  • Conventional statistical methods for group comparison require specific model assumptions.
  • Identifying the source of group differences is crucial for selecting appropriate statistical tests.
  • Model misspecification can lead to inaccurate conclusions in group difference analysis.

Purpose of the Study:

  • To propose a novel group comparison approach using machine learning classification algorithms without model specification.
  • To evaluate the performance of Support Vector Machines (SVMs) against conventional statistical tests.
  • To demonstrate the applicability of SVMs for detecting group differences in various data types and scenarios.

Main Methods:

  • Utilized classification algorithms, specifically Support Vector Machines (SVMs), for group comparison.
Keywords:
ClassifiersGroup comparisonIndependent validationK-fold cross validationSupport vector machine

Related Experiment Videos

  • Employed Independent Validation to evaluate classification accuracy against a binomial distribution.
  • Compared SVM performance (false-positive errors, statistical power) against t-tests, Levene's test, K-S test, Fisher's z-transformation, and MANOVA.

    • Main Results:

    • SVMs detected group differences irrespective of their origin (mean, variance, distribution, covariance).
    • SVMs demonstrated consistent statistical power across different conditions, particularly improving with multiple sources of difference and more variables.
    • When differences stemmed from a single source, SVM power was lower than the optimal conventional test, but superior with multiple sources.

    Conclusions:

    • Machine learning classification offers a flexible alternative or complement to conventional group comparison tests.
    • SVMs are applicable to diverse data without requiring sophisticated model specification, addressing limitations of traditional methods.
    • This approach empowers researchers to test two-sample data even when unsure of the appropriate statistical test or facing assumption violations.