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

Hazard function estimators: a simulation study.

K R Hess1, D M Serachitopol, B W Brown

  • 1Department of Biomathematics, Box 237, The University of Texas M. D. Anderson Cancer Center, 1515 Holcombe Blvd., Houston, Texas 77030-4095, USA. khess@mdanderson.org

Statistics in Medicine
|November 2, 1999
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

Safety and activity of vandetanib in combination with everolimus in patients with advanced solid tumors: a phase I study.

ESMO open·2021
Same author

A feasibility study of neoadjuvant talazoparib for operable breast cancer patients with a germline <i>BRCA</i> mutation demonstrates marked activity.

NPJ breast cancer·2017
Same author

Postoperative MRI Evaluation of a Radiofrequency Cordotomy Lesion for Intractable Cancer Pain.

AJNR. American journal of neuroradiology·2017
Same author

Florfenicol residues in rainbow trout after oral dosing in recirculating and flow-through culture systems.

Journal of aquatic animal health·2014
Same author

The prognostic significance of left ventricular ejection fraction in patients with advanced cancer treated in phase I clinical trials.

Annals of oncology : official journal of the European Society for Medical Oncology·2013
Same author

The prognostic impact of age in patients with triple-negative breast cancer.

Breast cancer research and treatment·2013

Kernel-based hazard estimation methods were compared using simulations. Locally optimal bandwidths with boundary correction significantly improved accuracy, outperforming existing algorithms for survival data analysis.

Area of Science:

  • Statistics
  • Survival Analysis
  • Non-parametric Estimation

Background:

  • Kernel-based methods are widely used for hazard function estimation.
  • While mathematical properties are known, statistical properties of these estimators require further investigation.
  • Existing methods often lack optimal bandwidth selection and boundary correction, impacting accuracy.

Purpose of the Study:

  • To review and compare statistical properties of kernel-based hazard estimators.
  • To evaluate the impact of boundary correction and locally optimal bandwidths.
  • To assess performance against existing algorithms for right-censored data.

Main Methods:

  • Computer simulations were employed to compare estimators.
  • Seven distributions, three censoring levels, four bandwidth types, and three boundary correction types were simulated.

Related Experiment Videos

  • A total of 504 experiments with 500 samples each were conducted.

    • Main Results:

    • Locally optimal bandwidths combined with boundary correction yielded substantial improvements.
    • Median relative improvement in mean square error reached 66% compared to basic estimators.
    • These advanced methods outperformed three existing algorithms by up to 80%.

    Conclusions:

    • Boundary correction and locally optimal bandwidths are crucial innovations for kernel hazard estimation.
    • The proposed methods offer superior accuracy and robustness for survival data analysis.
    • These findings provide guidance for selecting effective hazard estimation techniques.