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 hybrid SEM algorithm for high-dimensional unsupervised learning using a finite generalized Dirichlet mixture.

Nizar Bouguila1, Djemel Ziou

  • 1Concordia Institute for Information Systems Engineering (CIISE), Concordia University, Montréal, QC H3G IT7, Canada. bouguila@ciise.concordia.ca

IEEE Transactions on Image Processing : a Publication of the IEEE Signal Processing Society
|September 5, 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

Neural topic modeling on hyperspheres: Spherical representation learning with von Mises-Fisher mixtures.

Neural networks : the official journal of the International Neural Network Society·2026
Same author

HKANLP: Link Prediction With Hyperspherical Embeddings and Kolmogorov-Arnold Networks.

IEEE transactions on neural networks and learning systems·2025
Same author

Disentangled representation learning for multi-view clustering via von Mises-Fisher hyperspherical embedding.

Neural networks : the official journal of the International Neural Network Society·2025
Same author

Clustering and Interpretability of Residential Electricity Demand Profiles.

Sensors (Basel, Switzerland)·2025
Same author

Correlated Topic Modeling for Short Texts in Spherical Embedding Spaces.

IEEE transactions on pattern analysis and machine intelligence·2025
Same author

SAVE: Self-Attention on Visual Embedding for Zero-Shot Generic Object Counting.

Journal of imaging·2025
Same journal

Change-Prior-Guided Unsupervised Change Detection of Heterogeneous Remote Sensing Images.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
Same journal

AgonicDreamer: Enhancing Multi-View Consistency in Text-to-3D Generation via Rectified Score Distillation.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
Same journal

BiCM-Prompt: Bidirectional Cross-Modal Prompt Tuning for Class-Incremental Learning on Multisource Remote Sensing Images.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
Same journal

GoP-based Quality Enhancement on Video Compression.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
Same journal

Align then Tensorize: Multi-Level Consistent Anchor Graph Learning for Scalable Multi-View Clustering.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
Same journal

Beyond Fidelity: Diverse Image Synthesis via Retrieval-Augmented Diffusion.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
See all related articles

This study introduces a flexible generalized Dirichlet mixture model for unsupervised learning in high-dimensional data. The novel hybrid stochastic expectation maximization algorithm efficiently handles complex distributions and automatically determines the number of components.

Area of Science:

  • Statistics
  • Machine Learning
  • Data Science

Background:

  • Unsupervised learning of high-dimensional data presents challenges due to complexity and potential information loss with dimensionality reduction.
  • Existing mixture models may lack flexibility in approximating diverse data distributions (symmetric and asymmetric).

Purpose of the Study:

  • To develop and analyze a novel finite mixture model based on a generalized Dirichlet distribution for high-dimensional unsupervised learning.
  • To introduce a hybrid stochastic expectation maximization (HSEM) algorithm for parameter estimation and automatic component selection.
  • To demonstrate the model's utility in pattern recognition, image restoration, and texture image database summarization.

Main Methods:

  • Development of a generalized Dirichlet distribution offering a more flexible covariance structure than the standard Dirichlet distribution.

Related Experiment Videos

  • Proposal of a hybrid stochastic expectation maximization (HSEM) algorithm incorporating random component assignment and Newton-Raphson steps.
  • Implementation of an agglomerative term within HSEM for autonomous selection of the number of mixture components.

    • Main Results:

    • The generalized Dirichlet mixture model enables high-dimensional modeling without information loss from dimensionality reduction.
    • The HSEM algorithm effectively estimates model parameters and automatically determines the optimal number of components.
    • Successful application of the model to pattern recognition, image restoration, and texture image database summarization, including results on the Vistex database.

    Conclusions:

    • The generalized Dirichlet mixture model provides a powerful and practical tool for unsupervised learning in high-dimensional settings.
    • The HSEM algorithm offers an efficient and robust method for model training and component selection.
    • The model demonstrates significant potential for various computer vision and data analysis tasks.