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

Nonsmooth nonnegative matrix factorization (nsNMF).

Alberto Pascual-Montano1, J M Carazo, Kieko Kochi

  • 1Computer Architecture and System Engineering Department, Facultad de Ciencias Físicas, Universidad Complutense, Madrid, Spain. pascual@fis.ucm.es

IEEE Transactions on Pattern Analysis and Machine Intelligence
|March 11, 2006
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

AlignPCA-2D: PCA-reduced Euclidean vector alignment for 2D classification in cryo-EM.

Acta crystallographica. Section D, Structural biology·2026
Same author

Rational acquisition of laboratory equipment: an accurate mathematical model to estimate the trade-offs in shared and nonshared equipment.

Acta crystallographica. Section D, Structural biology·2026
Same author

PyCaret machine learning library with three preprocessing steps after eLORETA source estimation predicts Alzheimer's disease.

Neuroimage. Reports·2026
Same author

How many (distinguishable) classes can we identify in single-particle analysis?

Acta crystallographica. Section D, Structural biology·2025
Same author

Cortical electrical activity changes in healthy aging using EEG-eLORETA analysis.

Neuroimage. Reports·2025
Same author

Automatic detection of alignment errors in cryo-electron tomography.

Journal of structural biology·2024
Same journal

HardFlow: Hard-Constrained Sampling for Flow-Matching Models Via Trajectory Optimization.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Industrial Brain: Self-Evolving Neuro-Symbolic Autonomy with Causal Resilience for Cyber-Physical Systems.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Adaptive Hardness-Driven Dictionary Distillation for Incomplete Streaming View Clustering.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Mixture of Global and Local Experts with Diffusion Transformer for Controllable Face Generation.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Task-KV: Task-aware KV Cache Optimization via Semantic Differentiation of Attention Heads.

IEEE transactions on pattern analysis and machine intelligence·2026
Same journal

Achieving Text-based Person Retrieval with Any Granularity.

IEEE transactions on pattern analysis and machine intelligence·2026
See all related articles

We introduce nonsmooth nonnegative matrix factorization (nsNMF), a new method for analyzing data. This technique enhances data representation by finding localized patterns, improving interpretability and sparseness compared to traditional methods.

Area of Science:

  • Data analysis
  • Machine learning
  • Multivariate statistics

Background:

  • Nonnegative matrix factorization (NMF) is a common technique for data decomposition.
  • Classical NMF may not always yield easily interpretable or localized patterns.
  • There is a need for NMF variants that explicitly promote sparsity and part-based representations.

Purpose of the Study:

  • To propose a novel nonnegative matrix factorization model called nonsmooth nonnegative matrix factorization (nsNMF).
  • To develop a method that finds localized, part-based representations of nonnegative multivariate data.
  • To enhance the interpretability and sparseness of factorizations.

Main Methods:

  • Developed a new cost function for nonnegative matrix factorization that explicitly incorporates nonsmoothness.

Related Experiment Videos

  • Introduced a single parameter to control the degree of sparseness in the factorization.
  • Applied the nonsmooth nonnegative matrix factorization (nsNMF) model to various datasets.

    • Main Results:

    • The nsNMF model successfully generates basis and encoding vectors that represent the original data.
    • The method extracts highly localized patterns, leading to improved interpretability.
    • Compared to existing methods, nsNMF demonstrates advantages in data faithfulness and achieving high sparseness.

    Conclusions:

    • Nonsmooth nonnegative matrix factorization (nsNMF) offers a powerful alternative to classical NMF.
    • The proposed method enhances data interpretability through localized, part-based representations.
    • nsNMF effectively balances data representation fidelity with factor sparseness.