An Investigation and Interpretation of Selected Topics in Uncertainty Reasoning

Abstract

Incorporating techniques for coping with uncertainty in the decision support systems has proven to be a fertile environment for creative ideas. Representations of uncertainty abound and no representation can be said to be inherently incorrect. From a theoretical standpoint, a viable solution must be coherent and logically consistent. Probability theory demonstrates these characteristics while, as of yet, other methods do not. The purpose of this study was to investigate specific topics in uncertainty reasoning: 1)Probability ratio graphs as a representation of the probability model; 2) Dealing with missing information when system parameters are left unspecified; 3) Investigating the difference between probabilistic and causal independence; and, 4) Characterizing secondary uncertainty as spurious evidence and including it in the inference process. It was shown that probability ratio graphs are a viable method for representing uncertainty, and a method for representing independence with probability ratio graphs is presented. Assuming probabilistic independence for missing information is shown to have intuitive and computational benefits; also shown is that where secondary uncertainty is included in the inference process has great impact on the computational complexity of an inference process. Keywords: Bayes Theorem, Probability, Reasoning, Theses.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Dec 01, 1989
Accession Number
ADA215370

Entities

People

  • Scott E. Deakin

Organizations

  • Air Force Institute of Technology

Tags

Communities of Interest

  • Autonomy
  • Human Systems
  • Space

DTIC Thesaurus Topics

  • Artificial Intelligence
  • Bayesian Inference
  • Cognition
  • Computational Complexity
  • Computer Programming
  • Computer Programs
  • Computers
  • Delphi Method
  • Expert Systems
  • Fuzzy Sets
  • Mathematical Models
  • Operations Research
  • Probabilistic Models
  • Probability
  • Reasoning
  • Set Theory
  • Statistical Analysis

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Artificial Intelligence
  • Theoretical Analysis.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference