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

On visualization and aggregation of nearest neighbor classifiers.

Anil K Ghosh1, Probal Chaudhuri, C A Murthy

  • 1Theoretical Statistics and Mathematics Unit, Indian Statistical Institute, 203 B.T. Road, Kolkata-700 108, India. anilkghosh@rediffmail.com

IEEE Transactions on Pattern Analysis and Machine Intelligence
|October 22, 2005
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

Full maturation of in vitro Plasmodium falciparum oocysts using the AlgiMatrix 3D culture system.

Malaria journal·2024
Same author

Performance assessment of genomic island prediction tools with an improved version of Design-Island.

Computational biology and chemistry·2022
Same author

The fibrinolytic system enables the onset of <i>Plasmodium</i> infection in the mosquito vector and the mammalian host.

Science advances·2021
Same author

Purification and production of Plasmodium falciparum zygotes from in vitro culture using magnetic column and Percoll density gradient.

Malaria journal·2020
Same author

On Perfect Clustering of High Dimension, Low Sample Size Data.

IEEE transactions on pattern analysis and machine intelligence·2019
Same author

The Development of Whole Sporozoite Vaccines for <i>Plasmodium falciparum</i> Malaria.

Frontiers in immunology·2019

This study introduces a new approach to k-nearest neighbor classification, using a sequence of classifiers and a Bayesian measure of strength. This method improves classification accuracy over traditional single-k approaches.

Area of Science:

  • Statistical pattern recognition
  • Machine learning
  • Data mining

Background:

  • Nearest neighbor classification is a popular statistical pattern recognition method.
  • Determining the optimal neighborhood parameter k is a significant challenge.
  • Cross-validation is commonly used but may not capture data-specific optimal k.

Purpose of the Study:

  • To address the limitations of single-value k in nearest neighbor classification.
  • To propose and investigate a novel Bayesian measure of strength for class evidence.
  • To develop an improved classification method using a sequence of k-nearest neighbor classifiers.

Main Methods:

  • Utilizing a finite sequence of classifiers indexed by k.
  • Introducing and evaluating a Bayesian measure of strength.

Related Experiment Videos

  • Employing visual displays (shaded strips) for evidence visualization.
  • Proposing a weighted averaging technique for aggregating classifier results.

    • Main Results:

    • The proposed method demonstrates effective visualization of class evidence.
    • A Bayesian measure of strength provides a new metric for evidence.
    • Weighted averaging of multiple k-nearest neighbor classifiers improves performance.
    • The new approach outperforms using a single k value on benchmark datasets.

    Conclusions:

    • A sequence of k-nearest neighbor classifiers with a Bayesian measure of strength offers enhanced classification.
    • Visualizations aid in understanding evidence for different classes.
    • Weighted averaging provides a robust method for final decision-making.
    • The proposed method represents an advancement over traditional k-nearest neighbor techniques.