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

Macrostate data clustering.

Daniel Korenblum1, David Shalloway

  • 1Biophysics Program, Department of Molecular Biology and Genetics, Cornell University, Ithaca, NY 14853, USA.

Physical Review. E, Statistical, Nonlinear, and Soft Matter Physics
|June 6, 2003
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

A transit-amplifying progenitor with biphasic behavior contributes to epidermal renewal.

Development (Cambridge, England)·2024
Same author

Binary organization of epidermal basal domains highlights robustness to environmental exposure.

The EMBO journal·2022
Same author

Chromatin Modifiers Alter Recombination Between Divergent DNA Sequences.

Genetics·2019
Same author

Laplacian mixture modeling for network analysis and unsupervised learning on graphs.

PloS one·2018
Same author

Unfair competition governs the interaction of pCPI-17 with myosin phosphatase (PP1-MYPT1).

eLife·2017
Same author

Acetonitrile cluster solvation in a cryogenic ethane-methane-propane liquid: Implications for Titan lake chemistry.

The Journal of chemical physics·2017

This study introduces a novel nonhierarchical data clustering method inspired by stochastic systems. It effectively identifies fuzzy clusters representing metastable states, outperforming existing methods on challenging datasets.

Area of Science:

  • Computational science
  • Data analysis
  • Statistical mechanics

Background:

  • Traditional clustering methods often struggle with complex, overlapping, or fuzzy data structures.
  • Existing spectral partitioning methods may not adequately capture the dynamic or metastable nature of certain systems.

Purpose of the Study:

  • To develop a novel nonhierarchical data clustering algorithm.
  • To leverage principles from statistical mechanics, specifically dynamic coarse-graining, for data analysis.
  • To create a method that robustly identifies fuzzy clusters representing metastable states in diffusive systems.

Main Methods:

  • Developed a nonhierarchical clustering approach based on the dynamic coarse-graining analogy.
  • Analyzed the eigensystem of an interitem transition matrix to identify fuzzy clusters.

Related Experiment Videos

  • Employed a "minimum uncertainty criterion" for eigenvector-to-window function transformation.
  • Utilized eigenspectrum gap and cluster certainty for determining the optimal number of clusters.

    • Main Results:

    • The proposed method effectively identifies fuzzy clusters corresponding to metastable macroscopic states (macrostates).
    • Physically motivated fuzzy representation and uncertainty analysis provide a distinct advantage over spectral partitioning.
    • The macrostate data clustering approach successfully addressed test cases that challenge other clustering algorithms.

    Conclusions:

    • The developed macrostate data clustering method offers an effective approach for nonhierarchical clustering.
    • This physically grounded method provides robust identification of fuzzy clusters and their underlying system states.
    • The technique demonstrates superior performance on complex datasets and challenging clustering problems.