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

A multi-clustering fusion scheme for data partitioning.

Dimitrios S Frossyniotis1, Christos Pateritsas, Andreas Stafylopatis

  • 1National Technical University of Athens, School of Electrical and Computer Engineering, Zographou 157 80, Athens, Greece. dfros@cslab.ntua.gr

International Journal of Neural Systems
|November 10, 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

Localizing Perturbations in Pressurized Water Reactors Using One-Dimensional Deep Convolutional Neural Networks.

Sensors (Basel, Switzerland)·2022
Same author

Self-Organizing Hidden Markov Model Map (SOHMMM).

Neural networks : the official journal of the International Neural Network Society·2013
Same author

A review of decision support systems in telecare.

Journal of medical systems·2003
See all related articles

This study introduces a novel multi-clustering fusion method that combines multiple algorithm runs for stable data partitioning. This approach overcomes clustering instabilities and identifies the optimal number of clusters reliably.

Area of Science:

  • Data Science
  • Machine Learning
  • Computational Statistics

Background:

  • Clustering algorithms often suffer from instability due to initialization sensitivity.
  • Achieving a robust and consistent data partition remains a challenge in unsupervised learning.
  • Determining the optimal number of clusters is crucial for meaningful data interpretation.

Purpose of the Study:

  • To present a multi-clustering fusion method for enhanced data partitioning.
  • To overcome the inherent instabilities and initialization dependencies of single clustering runs.
  • To develop a procedure for reliably determining the optimal number of clusters.

Main Methods:

  • Combining results from multiple independent runs of a clustering algorithm.
  • Generating a common, stable data partition resistant to initialization variations.

Related Experiment Videos

  • Applying a fusion procedure to identify the optimal cluster count based on predefined criteria.

    • Main Results:

    • A distinct data partition unaffected by initialization randomness.
    • Mitigation of instabilities commonly observed in clustering methods.
    • A systematic approach to selecting the optimal number of clusters.

    Conclusions:

    • The proposed multi-clustering fusion method offers a robust solution for data partitioning.
    • This technique enhances the reliability and stability of clustering results.
    • It provides a reliable framework for determining the optimal number of clusters in datasets.