An Analysis of Robust and Efficient Priors Associated with a Finite Bayesian Model for Compliance Testing.

Abstract

This research examines the reliability and validity of the Finite Bayesian Procedure (FBP) model through an evaluation of robust and efficient prior probability distributions. The model, developed by James Godfrey and Richard Andrews, presents a different approach to compliance testing in auditing. This study utilizes small and moderate-sized populations, four population error rates, a fixed sample size, and four reliability levels. In addition, four expected error rates, based on a beta prior probability distribution and ranging from very low to high, combined with three variance levels and a uniform distribution, are used to evaluate the model. The results indicate that the model is adequately reliable and valid. However, the uniform distribution seems to perform best of all prior probability distributions tested. Moreover, tradeoffs between robustness, efficiency, and reliability seem a necessity when using the Finite Bayesian Procedure model. (Author)

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Sep 01, 1984
Accession Number
ADA148497

Entities

People

  • K. O'shaughnessy
  • S. Marino

Organizations

  • Air Force Institute of Technology

Tags

Communities of Interest

  • Cyber

DTIC Thesaurus Topics

  • Accounting
  • Accuracy
  • Air Force
  • Auditing
  • Bayes Theorem
  • Bayesian Inference
  • Bayesian Networks
  • Computational Science
  • Data Science
  • Models
  • Plastic Explosives
  • Probability
  • Probability Distributions
  • Reliability
  • Security
  • Statistical Sampling
  • Test And Evaluation

Readers

  • Computational Modeling and Simulation
  • Regression Analysis.
  • Statistical inference.

Technology Areas

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