Plain text is unavailable for this page.

An Introduction to Knowledge Engineering

S.L. Kendal and M. Creen An Introduction to Knowledge Engineering With 33 figures

S.L. Kendal School of Computing & Technology University of Sunderland Tyne and Wear UK M. Creen Learning Development Services University of Sunderland Tyne and Wear UK British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library Library of Congress Control Number: 2006925857 ISBN 10: 1-84628-475-9 Printed on acid-free paper ISBN 13: 978-1-84628-475-5 C Springer-Verlag London Limited 2007 Apart from any fair dealing for the purposes of research or private study, or criticism or review, as permitted under the Copyright, Designs and Patents Act 1988, this publication may only be reproduced, stored or transmitted, in any form or by any means, with the prior permission in writing of the publishers, or in the case of reprographic reproduction in accordance with the terms of licences issued by the Copyright Licensing Agency. Enquiries concerning reproduction outside those terms should be sent to the publishers. The use of registered names, trademarks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant laws and regulations and therefore free for general use. The publisher makes no representation, express or implied, with regard to the accuracy of the infor- mation contained in this book and cannot accept any legal responsibility or liability for any errors or omissions that may be made. Printed in the United States of America (TB/MVY) 9 8 7 6 5 4 3 2 1 Springer Science+Business Media springer.com

To my wife Janice, who is a better partner than I could wish for, and my daughter Cara, a gift from God. ���Simon Kendal To Lillian and Sholto���with love. ���Malcolm Creen

Foreword An Introduction to Knowledge Engineering presents a simple but detailed explo- ration of current and established work in the field of knowledge-based systems and related technologies. Its treatment of the increasing variety of such systems is designed to provide the reader with a substantial grounding in such technolo- gies as expert systems, neural networks, genetic algorithms, case-based reasoning systems, data mining, intelligent agents and the associated techniques and method- ologies. The material is reinforced by the inclusion of numerous activities that provide opportunities for the reader to engage in their own research and reflection as they progress through the book. In addition, self-assessment questions allow the student to check their own understanding of the concepts covered. The book will be suitable for both undergraduate and postgraduate students in computing science and related disciplines such as knowledge engineering, artificial intelligence, intelligent systems, cognitive neuroscience, robotics and cybernetics. vii

Contents Foreword vii 1 An Introduction to Knowledge Engineering............................... 1 Section 1: Data, Information and Knowledge ................................ 2 Section 2: Skills of a Knowledge Engineer ................................... 10 Section 3: An Introduction to Knowledge-Based Systems ................. 18 2 Types of Knowledge-Based Systems ......................................... 26 Section 1: Expert Systems........................................................ 27 Section 2: Neural Networks...................................................... 36 Section 3: Case-Based Reasoning............................................... 55 Section 4: Genetic Algorithms................................................... 66 Section 5: Intelligent Agents..................................................... 74 Section 6: Data Mining ........................................................... 83 3 Knowledge Acquisition.......................................................... 89 4 Knowledge Representation and Reasoning ................................ 108 Section 1: Using Knowledge..................................................... 109 Section 2: Logic, Rules and Representation .................................. 116 Section 3: Developing Rule-Based Systems .................................. 126 Section 4: Semantic Networks................................................... 140 Section 5: Frames .................................................................. 149 5 Expert System Shells, Environments and Languages ................... 159 Section 1: Expert System Shells................................................. 160 Section 2: Expert System Development Environments ..................... 165 Section 3: Use of AI Languages................................................. 168 ix

x Contents 6 Life Cycles and Methodologies................................................ 183 Section 1: The Need for Methodologies ....................................... 185 Section 2: Blackboard Architectures ........................................... 192 Section 3: Problem-Solving Methods .......................................... 199 Section 4: Knowledge Acquisition Design System (KADS)............... 209 Section 5: The Hybrid Methodology (HyM).................................. 218 Section 6: Building a Well-Structured Application Using Aion BRE.... 232 7 Uncertain Reasoning............................................................. 239 Section 1: Uncertainty and Expert Systems................................... 240 Section 2: Confidence Factors ................................................... 243 Section 3: Probabilistic Reasoning.............................................. 248 Section 4: Fuzzy Logic............................................................ 259 8 Hybrid Knowledge-Based Systems........................................... 270 Bibliography 283 Index 285

1 An Introduction to Knowledge Engineering Introduction This chapter introduces some of the key concepts in knowledge engineering. Al- most all of the topics are covered in summary form, and they will be explained in more detail in subsequent chapters. The chapter consists of three sections: 1. Data, information and knowledge 2. Skills of a knowledge engineer 3. An introduction to knowledge-based systems (KBSs). Objectives By the end of this chapter, you will be able to: define knowledge and explain its relationship to data and information distinguish between knowledge management and knowledge engineering explain the skills required of a knowledge engineer comment on the professionalism, methods and standards required of a knowledge engineer explain the difference between knowledge engineering and artificial intelligence define KBSs explain what a KBS can do explain the differences between human and computer processing state a brief definition of expert systems, neural networks, case-based reasoning, genetic algorithms, intelligent agents and data mining. 1

2 An Introduction to Knowledge Engineering SECTION 1: DATA, INFORMATION AND KNOWLEDGE Introduction This section defines knowledge and explains its relationship to data and informa- tion. Objectives By the end of this section you will be able to: develop a working definition of knowledge and describe its relationship to data and information. What Is Knowledge Engineering? ���Knowledge engineering is the process of developing knowledge based systems in any field, whether it be in the public or private sector, in commerce or in industry��� (Debenham, 1988). But what, precisely, is knowledge? What Is Knowledge? Knowledge is ���The explicit functional associations between items of information and/or data��� (Debenham, 1988). Data, Information and Knowledge What is data? Is it the same as information? Before we can attempt to understand what knowledge is, we should at least attempt to come closer to establishing exactly what data and information are.

An Introduction to Knowledge Engineering 3 Activity 1 The following activity introduces you to the concepts of data and information: 1. Read the following descriptions and definitions of ���data��� drawn from a variety of sources: Data (the plural of datum) are just raw facts (Long and Long, 1998). Data . . . are streams of raw facts representing events . . . before they have been arranged into a form that people can understand and use (Laudon and Laudon, 1998). Data is comprised of facts (Hayes, 1992). Recorded symbols (McNurlin and Sprague, 1998). 2. Make a note of any factors common to two or more of the descriptions. Feedback 1 You will have noticed that data is often spoken of as the same as ���facts������often ���raw��� and, in the first quotation, considered to move in a ���stream���. The final quo- tation from Hayes appears to look deeper in defining data more fundamentally as recorded symbols. Hayes actually goes on to insist that data are not facts and that treating them as such can produce ���innumerable perversions��� for example, in the form of propaganda or lies���which are still ���data���. You do not need to accept or reject any of the definitions you encounter���simply be aware that there are no universally accepted definitions of data. Similarly, in connection to the meaning of the term ���information���, we find that there are many attempts at definitions in the textbooks on information systems and information technology. In many ways the meanings of the words ���data��� and ���information��� only become clearer when we approach the differences between them. The following activity will help you to appreciate this. Activity 2 This activity introduces you to some definitions of information and its relation- ship to data. 1. Read the following definitions and descriptions of information. As in the last activity look for common denominators. That property of data which represents and measures effects of processing them (Hayes, 1992). By information we mean data that have been shaped into a form that is meaningful and useful to human beings (Laudon and Laudon, 1998).

4 An Introduction to Knowledge Engineering Information is data that have been collected and processed into a meaningful form. Simply, information is the meaning we give to accumulated facts (data) (Long and Long, 1998). Information is the emergent property which comes from processing data so that it is transformed into a structured whole (Harry, 1994). Information is data presented in a form that is meaningful to the recipient (Senn, 1990). Information is data in context (McNurlin and Sprague, 1998). Information is data endowed with relevance and purpose (Drucker, 1988). 2. Make a note of any similarities between the different descriptions. Feedback 2 You should have noted that information is commonly thought to be data, pro- cessed or transformed into a form or structure suitable for use by human beings. Such words as ���meaning���, ���meaningful���, ���useful��� and ���purpose��� are in evidence here. You may also have noted that information is considered a property of data. This implies that the former cannot exist without the latter. In the definitions of information you will have seen how the meaning of the word becomes clearer when the differences between it and data are considered. For example, whereas the ���rawness��� of data was emphasised earlier, informa- tion is considered to be some refinement of data for the purposes of human use. In addition, the words ���knowledge��� and ���communication��� have emerged as having a relationship to data and information. What is also worth emphasising at this point is that the interface between data and a human being���s interpretation of it is where information���determined by ���meaning������really emerges. The two terms are still often used interchangeably and no definition of either will apply in all the situations you might encounter. Knowledge In common language, the word knowledge is obviously related to information, but it is clear that they are not the same thing. So, how can we define knowledge in the same flexible way in which we have arrived at working definitions of information and data?

An Introduction to Knowledge Engineering 5 Activity 3 This activity extends your understanding of data and information. Look at the seven topics described briefly below. Which of them would you consider yourself as ���knowing���, and which would you consider yourself as having information about? (a) A second language in which you are fluent. (b) The content of a television news programme. (c) A close friend. (d) A company���s annual report. (e) Your close friend���s partner whom you have yet to meet. (f) The weather on the other side of the world. (g) The weather where you are now. Feedback 3 It is probable���but by no means certain���that you will have been inclined to consider items (a), (c) and (g) as things you can know about and the others as things for which you may have information. Note that the items that you would not describe yourself as possessing knowledge of could actually become known if circumstances were different, e.g. you might come to know your close friend���s partner. It is also worth noting that all of this depends on individual perceptions rather than measurable facts. You may only think you know your close friend. Simi- larly, your fluency in the second language will always be relatively poorer than that of a native speaker. Activity 4 This activity brings you closer to a definition by helping you highlight the differences between having information and possessing knowledge. What would you suggest is the primary characteristic that distinguishes the ���having information��� situations from the ���knowing��� situations you categorised in the previous activity? You will need to make sure that your description does not simply describe information or data, but must particularly take account of the former.

6 An Introduction to Knowledge Engineering Feedback 4 You should have been able to identify specific characteristics of knowledge that distinguish it from information similar to those highlighted in the following quotations. According to experts in the field, knowledge is: the result of the understanding of information (Hayes, 1992) the result of internalising information (Hayes, 1992) collected information about an area of concern (Senn, 1990) information with direction or intent���it facilitates a decision or an action (Zachman, 1987). Here it has become clear that knowledge is what someone has after understanding information. Often this understanding follows the development of a detailed or long-term relationship with the known person or thing. Such a process can often be accelerated when the need to use the information for a critical decision arises. This application of information to a decision or area of concern is particularly relevant in an organisational situation. However, it should be clear that data, information and knowledge are not static things in themselves but stages in the process of using data and transforming it into knowledge. On this basis they can be considered points along a continuum, moving from less to more usefulness to a human being, in much the same way as we all move along a continuum from young to old, but at no point can we be defined as either. Activity 5 Temperature and humidity readings are taken from various locations around one city. These readings are taken four times each day, and the results collated in a central location. The city is 12 miles in diameter. Readings taken on the periphery of the city can show, over time, how rain or adverse weather conditions start at one side of the city and move across to the other side. Details of adverse weather can be used to warn weather-sensitive activities such as cricket or tennis matches when to expect a break in play. Explain how a series of temperature and humidity readings can be transformed from data into knowledge.

An Introduction to Knowledge Engineering 7 Feedback 5 Data. Individual temperature and humidity readings, by themselves, are simply numbers, and therefore represent data. Information. Information on where the readings have been taken (e.g. at which point in the city) and at what time provides a trend to show how the temperature is currently changing. This information can be used by someone to make a decision. Knowledge. Knowing how the temperature and humidity are changing AND, knowing about how the weather can affect people living or working in the city will allow decisions to be made concerning the use of umbrellas, warm clothing, running a cricket or tennis match, etc. In this situation, two or more sets of information are related and can be processed to reach a decision. The movement from data to knowledge implies a shift from facts and figures to more abstract concepts, as shown in Figure 1.1. Value Concepts Data Information Knowledge Facts and figures The temperature outside is 5oC It is cold ��� put on a warm coat. Example It is cold outside. FIGURE 1.1. Data, information and knowledge. In other words: It is 5���C���data. It is cold���information. It is cold outside AND if it is cold you should wear a warm coat���knowledge. From a knowledge engineering perspective, it is useful to consider knowledge as something that can be expressed as a rule or useful to assist a decision, i.e., IF it is cold outside THEN wear a warm coat. The perceived value of data increases as it is transferred into knowledge, because the latter enables useful decisions to be made.