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

Time-oriented hierarchical method for computation of principal components using subspace learning algorithm.

Marko Jankovic1, Hidemitsu Ogawa

  • 1Control Department, Electrical Engineering Institute Nikola Tesla, Koste Glavinica 8a, 11000 Belgrade, Serbia, Serbia and Montenegro. elmarkoni@ieent.org

International Journal of Neural Systems
|December 14, 2004
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

Evolving Approaches to Bacterial Identification: A Review of Classical and Modern Techniques.

International journal of molecular sciences·2026
Same author

Fading on the beach: pigmentation plasticity adjusts coloration to substrate type in coastal Western spadefoot toads.

Current zoology·2025
Same author

Long-term salinity exposure reveals site-specific physiological and behavioural responses in coastal and inland toads.

The Journal of experimental biology·2025
Same author

A Potential Oncoprotective Role of Cytomegalovirus Against Breast Cancer: Worldwide Correlation and Survey of Evidence.

Diseases (Basel, Switzerland)·2025
Same author

A Comprehensive Overview of Antibacterial Agents for Combating Multidrug-Resistant Bacteria: The Current Landscape, Development, Future Opportunities, and Challenges.

Antibiotics (Basel, Switzerland)·2025
Same author

Nature's Arsenal: Uncovering Antibacterial Agents Against Antimicrobial Resistance.

Antibiotics (Basel, Switzerland)·2025
Same journal

Latent Space Projections and Atlases, a Cautionary Tale in Deep Neuroimaging using Autoencoders.

International journal of neural systems·2026
Same journal

Transformer-Based Anomaly Detection for Neurodegenerative Screening in MRI Images.

International journal of neural systems·2026
Same journal

Discrete Wavelet Convolution for Learnable Time-Frequency Representation with Application to Seizure Prediction.

International journal of neural systems·2026
Same journal

Automatic Seizure Detection using Hierarchical Spectral-Temporal Feature Learning with an Imbalance-Aware Transformer.

International journal of neural systems·2026
Same journal

Pyramid Vision Transformer-Enhanced Conformer Network for Epileptic Seizure Recognition Using MultiChannel EEG Signals.

International journal of neural systems·2026
Same journal

A Time-Frequency Decoupled Contrastive Learning Framework for Electroencephalography-Based Parkinson's Disease Diagnosis.

International journal of neural systems·2026
See all related articles

This study introduces a novel Principal Component Analysis (PCA) learning algorithm, modifying the Subspace Learning Algorithm (SLA) using a Time-Oriented Hierarchical Method (TOHM). This new approach enables neural networks to efficiently perform PCA by adapting basis vectors on two distinct time scales.

Area of Science:

  • Computational Neuroscience
  • Machine Learning
  • Statistical Data Analysis

Background:

  • Principal Component Analysis (PCA) and Principal Subspace Analysis (PSA) are fundamental techniques for dimensionality reduction and feature extraction.
  • Neural networks, utilizing gradient ascent/descent, can implement PSA and PCA, offering low-complexity solutions for analyzing correlations and updating eigenvectors.
  • Existing methods face challenges in efficiently adapting to slow changes in data correlations and updating principal components.

Purpose of the Study:

  • To propose a novel, fully homogeneous PCA learning algorithm for artificial neurons.
  • To adapt existing Principal Subspace Analysis (PSA) methods for Principal Component Analysis (PCA) using a hierarchical approach.
  • To enhance the capability of neural networks in performing PCA, particularly for tracking evolving data correlations.

Related Experiment Videos

Main Methods:

  • Modification of the Subspace Learning Algorithm (SLA), a prominent PSA learning algorithm.
  • Integration of the Time-Oriented Hierarchical Method (TOHM), employing two distinct time scales (fast and slow).
  • Development of a competitive mechanism among output neurons on the slower time scale to refine basis vectors towards principal eigenvectors.

Main Results:

  • A new PCA learning algorithm is presented that is fully homogeneous with respect to neurons.
  • The Time-Oriented Hierarchical Method (TOHM) effectively modifies PSA algorithms to perform PCA.
  • The proposed method facilitates the rotation of basis vectors within the principal subspace towards principal eigenvectors.

Conclusions:

  • The developed PCA learning algorithm offers a novel and efficient approach for neural network-based data analysis.
  • The Time-Oriented Hierarchical Method (TOHM) provides a versatile framework for transforming existing PSA neural network models into PCA models.
  • This research contributes to advancing neural network capabilities in feature extraction and data compression, especially in dynamic environments.