Nonparametric Methods for Hazard Rate Estimation from Right-Censored Samples.

Abstract

Nonparametric estimation of the hazard rate or failure rate is a frequent topic of investigation in the statistical literature because of its practical importance. Until quite recently, hazard rate estimation had been based on complete samples of independent identically distributed lifetimes. However, observations may be censored or truncated in many life testing situations. This occurs often in medical trials when the patients may enter treatment at different times and then either die from the disease under investigation or leave the study before its conclusion. A similar situation may occur in industrial life testing when items are removed from the test at random times for various reasons. It is of interest to be able to estimate nonparametrically the unknown hazard rate of the lifetime random variable from this type of data without ignoring or discarding the right-censored information. The purpose of this paper is to discuss nonparametric estimation of the hazard rate function for right-censored samples. The various types of estimators that have been proposed in the literature will be indicated and briefly discussed in Section 3. These include maximum likelihood estimators, kernel type estimators, Bayesian estimators, and histogram estimators. Due to their computational simplicity and other properties, the kernel-type hazard rate estimators will be emphasized. Results of Tanner (1983) and Tanner and Wong (1983, 1984) will be presented in Section 4 while the estimator considered by McNichols and Padgett (1981) will be discussed in Section 5.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1985
Accession Number
ADA159131

Entities

People

  • D. T. Mcnichols
  • William J. Padgett

Organizations

  • University of South Carolina

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Abstracts
  • Computational Science
  • Data Science
  • Estimators
  • Histograms
  • Information Science
  • Kernel Functions
  • Literature
  • Mathematics
  • Maximum Likelihood Estimation
  • Observation
  • Probability
  • Random Variables
  • South Carolina
  • Statistical Algorithms
  • Statistics
  • Theorems

Fields of Study

  • Mathematics

Readers

  • Educational Psychology
  • Statistical inference.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference