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

Improving Cox survival analysis with a neural-Bayesian approach.

Bart Bakker1, Tom Heskes, Jan Neijt

  • 1Theoretical Foundation SNN Laboratory, University of Nijmegen, Geert Grooteplein 21, 6525 EZ Nijmegen, The Netherlands. bart.bakker@philips.com

Statistics in Medicine
|September 8, 2004
PubMed
Summary

This study enhances Cox survival analysis using a neural Bayesian framework with priors. Bayesian methods offer more reliable predictions, especially for small datasets, outperforming traditional Cox 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

Feasibility to infer accurate fibrinogen concentrations from thrombin generation, prothrombin time, or Reptilase time waveforms.

Research and practice in thrombosis and haemostasis·2026
Same author

Structural equation modeling to explore putative causal factors for chronic fatigue in childhood cancer survivors: a DCCSS LATER study.

Journal of cancer survivorship : research and practice·2025
Same author

Long-term sequelae of SARS-CoV-2 two years following infection: exploring the interplay of biological, psychological, and social factors.

Psychological medicine·2024
Same author

Effect of Early Levodopa Treatment on Mortality in People with Parkinson's Disease.

Movement disorders clinical practice·2024
Same author

How to evaluate uncertainty estimates in machine learning for regression?

Neural networks : the official journal of the International Neural Network Society·2024
Same author

The effect of cardiovascular risk on disease progression in <i>de novo</i> Parkinson's disease patients: An observational analysis.

Frontiers in neurology·2023

Area of Science:

  • Biostatistics
  • Machine Learning
  • Survival Analysis

Background:

  • Traditional Cox survival analysis has limitations, particularly with high-dimensional data.
  • Improving predictive accuracy in survival analysis is crucial for various scientific fields.

Purpose of the Study:

  • To enhance Cox survival analysis using a neural Bayesian framework.
  • To demonstrate the superiority of Bayesian methods for survival prediction, especially with limited data.

Main Methods:

  • Integration of sensible priors into a neural Bayesian framework.
  • Approximation techniques for intractable posteriors: Hybrid Markov Chain Monte Carlo, variational methods, and Laplace approximation.
  • Development of an algorithm for input selection using Bayesian posterior and p-value estimation.

Related Experiment Videos

Main Results:

  • Bayesian framework provides more reliable predictions than traditional Cox analysis, particularly for small datasets.
  • Variational methods and Laplace approximation are practical approaches for Bayesian analysis.
  • Input selection via Bayesian methods significantly improves predictive performance over classical Cox analysis.

Conclusions:

  • Neural Bayesian framework offers a robust improvement over traditional Cox survival analysis.
  • Bayesian approaches, especially variational or Laplace, enhance predictive accuracy and handle high-dimensional inputs effectively.
  • The proposed method provides better predictions and input relevance assessment compared to classical Cox analysis.