Axiomatic Derivation of the Principle of Maximum Entropy and the Principle of Minimum Cross-Entropy.

Abstract

We prove that, in a well-defined sense, Jaynes's principle of maximum entropy and Kullback's principle of minimum cross-entropy (minimum directed divergence) provide uniquely correct, general methods of inductive inference when new information is given in the form of expected values. Previous justifications rely heavily on intuitive arguments and on the properties of entropy and cross-entropy as information measures. Our approach assumes that reasonable methods of inductive inference should lead to consistent results whenever there are different ways of taking the same information into account --- for example, in different coordinate systems. We formalize this requirement as four consistency axioms stated in terms of an abstract information operator; the axioms make no reference to information measures. We establish this result both directly and as a special case (uniform priors) of an analogous, more general result for the principle of minimum cross-entropy. We obtain results both for continuous probability densities and for discrete distributions.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 15, 1978
Accession Number
ADA063120

Entities

People

  • J. E. Shore
  • Wayne Johnson

Organizations

  • United States Naval Research Laboratory

Tags

Communities of Interest

  • C4I

DTIC Thesaurus Topics

  • Computer Programs
  • Computers
  • Convex Sets
  • Coordinate Systems
  • Differential Equations
  • Discrete Distribution
  • Equations
  • Inequalities
  • Information Theory
  • Invariance
  • Military Research
  • Notation
  • Probability
  • Probability Distributions
  • Reliability
  • Statistical Mechanics
  • Theorems

Readers

  • Artificial Intelligence
  • Mathematical Modeling and Probability Theory.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference
  • AI & ML - Machine Learning Algorithms