MUTUAL UNCERTAINTY AND PRIOR PROBABILITIES,

Abstract

The paper gives a general method of determining the prior probability function of a variable Lambda using only the prior knowledge which comprises the measurement scheme relating the observables X(1),X(2),... to Lambda, the range of Lambda and the additional knowledge such as the bounds on the expected value of Lambda. The essential idea is that one should not be able to draw any conclusions on the observable random vectors X(1),X(2),... which are not explicitly permitted by the prior knowledge. This leads us to the principle of Maximum Mutual Uncertainty for the determination of the prior probability function in all the cases. In particular, one can derive a rather simple formula for the prior density of a parameter. It is shown that the principle of Maximum Entropy developed by Jaynes for the determination of the prior density of the observable random variable is a special case of the principle of Maximum Mutual Uncertainty. A number of examples have been given to demonstrate the power of the method. (Author)

Document Details

Document Type
Technical Report
Publication Date
Jul 01, 1969
Accession Number
AD0695819

Entities

People

  • Rangasami L. Kashyap

Organizations

  • Purdue University

Tags

DTIC Thesaurus Topics

  • Mathematics
  • Measurement
  • Probability
  • Random Variables
  • Uncertainty

Fields of Study

  • Mathematics

Readers

  • Quantum spin resonance or Electron Paramagnetic Resonance spectroscopy.
  • Regression Analysis.
  • Systems Analysis and Design