Neyman-Pearson and Bayes interval estimates for sampling by attributes
This paper compares confidence intervals for single and multistage sampling schemes with Bayesian interval estimates obtained with a uniform prior distribution. Examples are presented in graphical form for sampling by attributes from an infinite population, or from a finite population with replacement. A general proof is given that the Neyman-Pearson confidence level associated with a confidence interval for the binomial parameter p will be no greater than the Bayesian confidence level calculated using a uniform prior distribution. A demonstration is provided for a fact published earlier, viz., that the Bayesian prior distribution can be selected so as to provide equality between one-sided Neyman-Pearson and Bayesian confidence bounds. Applications to EMP analysis are discussed in the final section.
- Research Organization:
- Booz Allen and Hamilton Inc., 2309 Renard Place, S.E., Suite 301, Albuquerque, NM 87106
- OSTI ID:
- 5844513
- Journal Information:
- IEEE Trans. Nucl. Sci.; (United States), Vol. NS-31:6
- Country of Publication:
- United States
- Language:
- English
Similar Records
Distributed decision fusion using empirical estimation
Confidence limits for parameters of Poisson and binomial distributions
Related Subjects
AIRCRAFT
PHYSICAL RADIATION EFFECTS
DATA TRANSMISSION SYSTEMS
MISSILES
ELECTROMAGNETIC PULSES
GRAPHS
MATHEMATICAL MODELS
PROBABILITY
SAMPLING
STRESSES
ELECTROMAGNETIC RADIATION
PULSES
RADIATION EFFECTS
RADIATIONS
440200* - Radiation Effects on Instrument Components
Instruments
or Electronic Systems