Engineering Risk and Hazard Assessment

Volume II

Abraham Kandel

• Publisher: CRC Press
• Year: 2018
• Print ISBN:
  9781315892610
• eBook ISBN:
  9781351088619
• Topic: Technology & Engineering
• Subtopic: Civil Engineering
• Index: Technology & Engineering

The volumes deal with the newly emerging field ofRisk and Hazard Assessment and its application to science and engineering. These volumes deal with issues such as short-and long-term hazards, setting priorities in safety, fault analysis for process plants, hazard identification and safety assessment of human- robot systems, plant fault diagnoses expert systems, knowledge based diagnostic systems, fault tree analysis, modelling of computer security systems for risk and reliability analysis, risk analysis of fatigue failure, fault evaluation of complex system, probabilistic risk analysis, and expert systems for fault detection. This volume will provide the reader not only with valuable conceptual and technical information but also with a better view of the field, its problems, accomplishments, and future potentials

Chapter 1

Measures of Accountability and Credibility in Knowledge Based Diagnosis Systems

Ronald R. Yager

TABLE OF CONTENTS

• I. Introduction
• II. Measures of Accountability
• III. Generalization of Accountability
• IV. Measure of Credibility
• V. Conclusion
• Acknowledgments
• References

I. Introduction

Many applications of knowledge based systems involve the determination of some underlying incidents based upon the observations of some set of symptoms. The classical example of such a system is a medical diagnosis system. Assume E is our set of symptoms and A is a hypothesized set of causes. To determine how good A is as a diagnosis for the sympton set E, we must measure the degree A accounts for the symptoms manifested in E. We call this a measure of accountability of A for E. However, this measure of accountability in and of itself is not a good indication of the quality of the diagnosis set A. For it is obvious that one can always increase the size of A to improve the accountability but without forming a better diagnosis. What is needed is an additional measure which we shall call the credibility of A given E. This measure of credibility is an indicator of the degree to which the symptoms which must be apparent if the diagnosis is A are contained in the set E. Thus the combination, in particular the conjunction, of these two measures provide an indication of the degree to which A is a good diagnosis in the light of the known evidence E.

In this paper we are primarily concerned with the development of formulations for these two measures. We note that the formulations developed allow for the possibility of multiple causes.

II. Measures of Accountability

In this section we introduce a measure of accountability which can be used to indicate the degree to whic...

Table of contents

• Cover
• Title Page
• Copyright Page
• Table of Contents
• Chapter 1 Measures of Accountability and Credibility in Knowledge Based Diagnosis Systems
• Chapter 2 The Modeling of Computer Security Systems Using Fuzzy Set Theory
• Chapter 3 Structure Modeling of Process Systems for Risk and Reliability Analysis
• Chapter 4 Risk Analysis of Fatigue Failure of Highway Bridges
• Chapter 5 Modularization Methods for Evaluating Fault Trees of Complex Technical Systems
• Chapter 6 Utilizing Probabilistic Risk Analyses (Pras) In Decision Support Systems
• Chapter 7 Survey of Expert Systems for Fault Detection, Test Generation, and Maintenance
• Index