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

Projective ART for clustering data sets in high dimensional spaces.

Yongqiang Cao1, Jianhong Wu

  • 1Department of Mathematics and Statistics, York University, Toronto, Ontario, Canada.

Neural Networks : the Official Journal of the International Neural Network Society
|April 18, 2002
PubMed
Summary

A novel neural network architecture, Projected ART (PART), efficiently finds projected clusters in high-dimensional data. Its selective output signaling mechanism overcomes sparsity, focusing on relevant dimensions for improved data mining.

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

Pre- and postoperative membranous urethral length on multiparametric magnetic resonance imaging predict stage-specific 0-pad continence recovery after robot-assisted radical prostatectomy.

Translational andrology and urology·2026
Same author

A Delay Differential Model for Antiviral Treatment and Rebound Dynamics.

Mathematical biosciences·2026
Same author

Synergistic effects of phosphatidylcholine and cholesterol on lipid droplet interfacial structure and in vitro digestibility of simulated breast milk fat emulsion.

Food research international (Ottawa, Ont.)·2026
Same author

Clarifying the "component-structure-function" mechanism in processed mozzarella cheese: the mechanistic role of fat, protein, and water interactions on quality and texture.

Food chemistry·2026
Same author

Variations in Ecological Locations Induce Soybean Seed Wrinkles by Disrupting Source-Sink Relationship and Energy Metabolism at the Grain-Filling Stage.

Plants (Basel, Switzerland)·2026
Same author

Exploring memory effects: Sparse identification in vector-borne diseases.

Proceedings of the National Academy of Sciences of the United States of America·2026

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Data Mining

Background:

  • High-dimensional data presents challenges for clustering due to sparsity.
  • Existing algorithms like Fuzzy ART and PROCLUS have limitations in handling such data.
  • Adaptive Resonance Theory (ART) provides a foundation for neural network-based clustering.

Purpose of the Study:

  • To introduce a new neural network architecture, Projected ART (PART), for projected clustering.
  • To address the sparsity issue in high-dimensional data mining applications.
  • To improve the efficiency and focus of clustering algorithms in high-dimensional spaces.

Main Methods:

  • Developed a novel neural network architecture, PART, based on ART.
  • Incorporated a selective output signaling mechanism to manage data sparsity.

Related Experiment Videos

  • Evaluated PART using illustrative examples, simulations on synthetic high-dimensional data, and comparisons with Fuzzy ART and PROCLUS.

    • Main Results:

    • PART successfully identifies projected clusters in high-dimensional data.
    • The selective output signaling mechanism effectively focuses on informative dimensions.
    • Simulations demonstrate PART's competitive performance against existing methods.

    Conclusions:

    • PART offers an effective solution for projected clustering in high-dimensional datasets.
    • The proposed architecture and selective signaling mechanism enhance data mining capabilities.
    • PART shows promise for applications dealing with sparse, high-dimensional data.