ESTIMATING CUMULATIVE PROBABILITY FROM AGGREGATED TRUNCATED DATA

Abstract

Cumulative survival, failure, or detection probabilities cannot, in general, be precisely estimated from truncated samples if only data grouped in sucessive time internals is available. Mathematical models of the failure rate and abort rate within the time intervals are postulated from which estimates may be obtained from grouped data when the models are valid. An easily calculable approximation formula can be used in the earlier time intervals where the sample size is relatively large. This can provide data for verifying or rejecting a given model prior to making calculations in later intervals where the smaller sample size would otherwise diminish the reliability of the resulting probabilities.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jul 06, 1964
Accession Number
AD0447677

Entities

People

  • H. Spitz

Organizations

  • Center for Naval Analyses

Tags

DTIC Thesaurus Topics

  • Acquisition
  • Contracts
  • Detection
  • Equations
  • Failure Mode And Effect Analysis
  • Government Procurement
  • Governments
  • Intervals
  • Mathematical Models
  • Models
  • Observation
  • Probability
  • Simultaneous Equations
  • Survival
  • Test And Evaluation
  • Time Intervals

Readers

  • Regression Analysis.