AN APPLICATION OF BAYES THEOREM AS A HYPOTHESIS-SELECTION AID IN A COMPLEX INFORMATION-PROCESSING SYSTEM.

Abstract

The first of a series of experiments investigating the value of automated hypothesis-evaluation aids in multimanmachine systems devoted to assessing or diagnosing threat is described. In the experiment, an eight-man team evaluated threats posed by a hypothetical aggressor. The team made these evaluations on the basis of intelligence information gathered during simulated reconnaissance overflights of aggressor's territory. The primary output of the threat-evaluation team was the commanding officer's posterior probabilities estimates as to aggressor's most likely hostile strategies. During half of the experimental trials, the commander had access to computer-produced posterior probabilities based upon a modification of the Bayes Theorem. The major experimental issue was whether or not these would aid the commander in his hypothesis evaluation. Also investigated was the effect of data-processing load upon system operation. Although some improvement in the posterior probabilities estimates resulted from the commander's having access to the hypothesis-evaluation aid and this improvement became more pronounced as system load increased, the main-order effect of access to the aid was not found to be statistically significant. (Author)

Document Details

Document Type
Technical Report
Publication Date
Aug 01, 1964
Accession Number
AD0607256

Entities

People

  • Davis A. Schum
  • George E. Briggs
  • Jack F. Southard

Organizations

  • Ohio State University

Tags

DTIC Thesaurus Topics

  • Bayes Theorem
  • Computers
  • Data Processing
  • Information Processing
  • Information Systems
  • Overflight
  • Probability
  • Reconnaissance
  • Test And Evaluation
  • Theorems
  • Threat Evaluation
  • Threats

Readers

  • Statistical inference.
  • Team-Based Human-Centered Cognitive Task Decision Making and Information Performance.