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

How many samples are needed to build a classifier: a general sequential approach.

Wenjiang J Fu1, Edward R Dougherty, Bani Mallick

  • 1Department of Statistics, Texas A&M University, 447 Blocker Building, College Station, TX 77843, USA. wfu@stat.tamu.edu

Bioinformatics (Oxford, England)
|August 7, 2004
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

Nonparametric Density Estimation of a Long-Term Trend from Repeated Semicontinuous Data.

Journal of the American Statistical Association·2026
Same author

Generative AI-assisted Bayesian-frequentist Hybrid Inference in Single-cell RNA Sequencing Analysis for Genes Associated with Alzheimer's Disease.

medRxiv : the preprint server for health sciences·2026
Same author

Application of the Total Nutrient Index as a Precision Nutrition Tool to Address Dietary Recommendations Across the Life Course.

Journal of the Academy of Nutrition and Dietetics·2026
Same author

LILRB4 regulates circadian disruption-induced mammary tumorigenesis via non-canonical WNT signaling pathway.

Oncogene·2025
Same author

Utilizing the effectiveness of phage cocktail to combat Shigella and Salmonella infections and their polymicrobial biofilm control activity.

BMC microbiology·2025
Same author

Valid and efficient inference for nonparametric variable importance in two-phase studies.

Biometrics·2025

This study introduces a sequential classification method that updates diagnostic rules as new patient data becomes available. This approach optimizes classifier accuracy while minimizing study costs through early stopping criteria.

Area of Science:

  • Biostatistics
  • Machine Learning
  • Bioinformatics

Background:

  • Traditional classifier design relies on fixed sample sizes, often limited by practical constraints like cost and time.
  • Limited sample sizes can hinder the development of accurate diagnostic classifiers.

Purpose of the Study:

  • To develop a sequential classification procedure that optimizes classifier performance while minimizing patient sample size.
  • To provide a cost-effective and efficient approach for classifier design in diagnostic studies.

Main Methods:

  • A sequential classification procedure was developed and analyzed using the martingale central limit theorem.
  • The method involves updating classification rules at each step and incorporating stopping criteria.
  • Simulation studies and analysis of microarray data were used to evaluate the procedure.

Related Experiment Videos

Main Results:

  • The sequential procedure effectively updates classification rules without relying on external data distributions.
  • Stopping criteria at each step allow for significant cost reduction through early termination of studies.
  • The method is versatile, applicable to any classification rule, including feature selection and extraction.

Conclusions:

  • The sequential classification approach offers an efficient and adaptable method for developing diagnostic classifiers.
  • This procedure can lead to substantial cost savings in research studies by minimizing sample size requirements.
  • The R-code for the sequential stopping rule is publicly available for implementation.