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

Minimizing model fitting objectives that contain spurious local minima by bootstrap restarting.

S N Wood1

  • 1The Mathematical Institute, University of St. Andrews, Fife, UK. snw@st-and.ac.uk

Biometrics
|March 17, 2001
PubMed
Summary

This study introduces a novel bootstrapping technique to improve nonlinear model fitting by perturbing objective functions. This method helps minimization algorithms escape local minima, enhancing convergence for complex data analysis.

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

Habitat structure and population persistence in an experimental community.

Nature·2001
Same author

Space, time and persistence of virulent pathogens.

Proceedings. Biological sciences·1996
Same author

Obtaining birth and mortality patterns from structured population trajectories.

Ecological monographs·1994
Same author

Spline models of biological population dynamics: how to estimate mortality rates for stage structured populations with dimorphic life histories.

IMA journal of mathematics applied in medicine and biology·1994
Same author

Factors influencing occupational back pain in Bedfordshire.

Spine·1980
Same author

Rural sanitation.

Canadian journal of public health = Revue canadienne de sante publique·1948

Area of Science:

  • Computational Statistics
  • Nonlinear Model Fitting
  • Data Science

Background:

  • Objective functions in nonlinear model fitting frequently present local minima, which can trap standard minimization algorithms.
  • Current methods rely on stochastic approaches to escape these local minima, treating the objective function as static.

Purpose of the Study:

  • To propose a simple, effective method for enhancing nonlinear minimization by stochastically perturbing the objective function.
  • To improve the ability of algorithms to avoid getting stuck in insignificant local minima during model fitting.

Main Methods:

  • A novel approach involving stochastic perturbation of the objective function via data bootstrapping.
  • Alternating minimizations between bootstrap-generated objective functions and the original objective function.

Related Experiment Videos

  • Utilizing parameter values from previous minimizations to initiate subsequent ones.

    • Main Results:

    • Demonstrated effectiveness in a case study fitting a nonlinear population dynamic model.
    • The bootstrapping method shows promise in overcoming local minima trapping.
    • Comparison with simulated annealing indicates competitive or superior performance.

    Conclusions:

    • The proposed data bootstrapping technique offers a practical enhancement for conventional nonstochastic fitting methods.
    • This approach effectively addresses the challenge of local minima in nonlinear model optimization.
    • Further investigation into convergence diagnostics is recommended for robust application.