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

Efficient methods for estimating constrained parameters with applications to lasso logistic regression.

Guo-Liang Tian1, Man-Lai Tang, Hong-Bin Fang

  • 1Division of Biostatistics, University of Maryland Greenebaum Cancer Center, 10 South Pine Street, MSTF Suite 261, Baltimore, Maryland 21201, U.S.A.

Computational Statistics & Data Analysis
|April 30, 2008
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

Niaoduqing particles ameliorated tubulointerstitial fibrosis by suppressing IκB/NF-κB signalling pathway via inhibiting host- and gut microbiota-mediated tryptophan co-metabolism.

Microbiological research·2026
Same author

Spatially Correlated Analysis of Infectious Disease Outcomes Based on Bayesian Functional Hierarchical Models.

Statistics in medicine·2026
Same author

Sesquiterpenoids from the fruits of Alpinia oxyphylla showing renal protective activity through NF-κB/IκBα and TGF-β1/Smad signaling pathways.

Phytochemistry·2026
Same author

Generalizations of the quadratic bound optimization principle.

Proceedings of the National Academy of Sciences of the United States of America·2026
Same author

Interruption of cross-communication pathways alters the immune cell signature of pancreatic cancer and decreases tumor growth.

Cell communication and signaling : CCS·2026
Same author

Functional varying-coefficient Cox model and its application.

Statistical methods in medical research·2026
Same journal

Modeling Disease-specific Survival in Observational Studies with Missing Cause of Death: Leveraging Information from Clinical Trial Data.

Computational statistics & data analysis·2026
Same journal

A simultaneous confidence-bounded true discovery proportion perspective on localizing differences in smooth terms in regression models.

Computational statistics & data analysis·2025
Same journal

MIXANDMIX: numerical techniques for the computation of empirical spectral distributions of population mixtures.

Computational statistics & data analysis·2024
Same journal

Locally sparse quantile estimation for a partially functional interaction model.

Computational statistics & data analysis·2024
Same journal

Flexible Regularized Estimation in High-Dimensional Mixed Membership Models.

Computational statistics & data analysis·2024
Same journal

GPU Accelerated Estimation of a Shared Random Effect Joint Model for Dynamic Prediction.

Computational statistics & data analysis·2024
See all related articles

We developed a new Quadratic Lower-Bound (QLB) algorithm for constrained logistic regression and a faster pseudo-Newton method. These methods improve optimization for logistic, multinomial logistic, and Cox models, enhancing statistical analysis.

Area of Science:

  • Statistics
  • Computational Statistics
  • Optimization

Background:

  • Fitting logistic regression models with restricted parameters presents computational challenges.
  • Existing algorithms like EM-type methods are not always applicable, and penalized regression convergence is not always guaranteed.
  • Efficient optimization is crucial for statistical modeling and inference.

Purpose of the Study:

  • To develop a novel Quadratic Lower-Bound (QLB) algorithm for optimization with box or linear inequality constraints.
  • To generalize the QLB algorithm for penalized regression, including non-differentiable penalty functions (e.g., lasso logistic regression).
  • To introduce a pseudo-Newton method for accelerated convergence in constrained optimization problems.

Main Methods:

  • Development of a Quadratic Lower-Bound (QLB) algorithm with a focus on the fastest variant corresponding to the smallest global majorization matrix.

Related Experiment Videos

  • Generalization of the QLB algorithm to handle penalized regression problems with potentially non-differentiable penalty functions.
  • Introduction of a pseudo-Newton method by relaxing the ascent requirement of the QLB algorithm for faster convergence.

    • Main Results:

    • The proposed QLB algorithm ensures monotonic convergence and is suitable for logistic, multinomial logistic, and Cox's proportional hazards models.
    • The generalized QLB algorithm offers an alternative for lasso logistic regression, addressing convergence issues of existing methods.
    • The pseudo-Newton method demonstrates significant speed improvements (up to 71x in CPU time, 107x in iterations) over the fastest QLB, enabling feasible bootstrap variance estimation.

    Conclusions:

    • The developed QLB and pseudo-Newton algorithms provide efficient and robust solutions for constrained and penalized logistic regression.
    • These methods enhance the applicability and reliability of statistical modeling for complex datasets.
    • The accelerated convergence of the pseudo-Newton method facilitates advanced statistical inference techniques like bootstrap variance estimation.