CONTENTS
PART 1 INTRODUCTION
CHAPTER 1 Introduction = 3
1.1 Expert Systems and Neural Networks as Qualitative Tools = 4
1.1.1. Quantitative Method as Tools for Analysis and Decision = 9
1.1.2. Can Quantitative Methods Address All Problems? = 10
1.1.3. Qualitative Nature of Expert Systems and Neural Networks = 12
1.1.4. Machine Intelligence = 16
1.2 A Brief History of Artificial Intelligence = 16
1.3 A Brief History of Neural Networks = 17
CHAPTER 2 Why Are Expert Systems and Neural Networks Needed? = 25
2.1 Applications of Expert Systems and Neural Networks = 27
2.1.1 Applications of Expert Systems = 28
2.1.2 Applications of Neural Networks = 30
2.2 Economics of Expert Systems and Neural Network Systems = 32
2.2.1. Technology as Impetus for Progress = 32
2.2.2. Expert Systems and Neural Networks as Productivity Tools = 38
2.2.3. Features of Expert Systems and Neural Networks as Productivity Tools = 40
2.2.4. Combination of Quantitative and Qualitative Tools = 40
2.3 The Synergy of Conventional and Intelligent Systems = 41
2.3.1. Synergy of Expert Systems and Database Systems = 41
2.3.2. Synergy of Expert Systems and Statistics = 42
2.3.3. Synergy of Neural Networks and Statistics = 43
2.3.4. Synergy of Decision Support Systems Tools with Expert Systems and Neutral Networks = 43
2.4 An Intergreted Approach to Expert Systems and Neural Networks = 44
2.5 Issues in Artificial Intelligence and Neural Networks = 46
2.5.1. The Criteria for Measuring Machine Intelligence = 46
2.5.2. Algorithmics in Quantitative Methods vs. Heuristics in Qualitative Methods = 48
2.5.3. The Debate over Machine Intelligence = 49
PART Ⅱ THE THEORETICAL FOUNDATION OF EXPERT SYSTEMS
CHAPTER 3 Knowledge Representation Based on Logic = 63
3.1 Structure of an Expert System = 66
3.1.1. Domain Knowledge = 66
3.1.2. Knowledge Base = 67
3.1.3. Human Component = 67
3.1.4. Expert System Software = 68
3.2 Logic-based knowledge Representation = 73
3.2.1. Rule-based Representation = 73
3.2.2. Logic as the Foundation for Knowledge Representation = 76
3.3 Propositional Logic = 79
3.3.1. Use of Connectives = 80
3.3.2. Truth Tables for Connectives = 80
3.3.3. Establishing the Truth Value of a Statement Form = 84
3.3.4. Tautology and Contradiction = 85
3.3.5. Truth Functions(Optional) = 86
3.4 Propositional Calculus = 87
3.5 Predicate Logic(Optional) = 88
3.5.1. Predicates(Optional) = 89
3.5.2. Quantifiers(Optional) = 90
3.5.3. Bound and Free Variables and Quantification(Optional) = 91
3.5.4. Relation of Quantifiers and Connectives(Optional) = 91
3.5.5. Multiple Quantifiers(Optional) = 92
3.6 Predicate Calculus = 92
3.7 Knowledge Representation for a Mortgage Loan Expert System = 93
3.7.1. Mortgage Loan Case = 93
3.7.2. Knowledge Base Represented in Rule-based Method = 94
3.7.3. Knowledge Base Represented in Predicate Method(Optional) = 95
CHAPTER 4 Inference and Konwledge Processing = 105
4.1 Reasoning Method = 107
4.2 Deductive Reasoning in Expert Systems = 108
4.3 Single Inference in Deductive Reasoning = 110
4.3.1. Inference in Propositional Logic and Calculus = 110
4.3.2. Inference in Predicate Calculus(Optional) = 115
4.3.3. Unification(Optional) = 116
4.3.4. Resolution(Optional) = 118
4.4 Multiple Inference in Deductive Reasoning = 124
4.4.1. Graphs, Trees, and the And / Or Graph = 124
4.4.2. Backward and Forward Chaining = 126
4.4.3. Search Methods : Depth-first and Breadth-first = 131
4.4.4. Other Heuristics in Expert Systems = 132
4.4.5. Shallow and Deep Reasoning = 134
4.5 Inductive Reasoning in Expert Systems = 134
4.5.1. Decision Trees = 135
4.5.2. ID3 = 137
4.5.3. Case-based Reasoning and Reasoning by Analogy = 139
PARTⅢ PRACTICAL ASPECTS IN APPLYING EXPERT SYSTEMS
CHAPTER 5 Deductive Reasoning Tools and LEVEL5 = 149
5.1 LEVEL5 = 151
5.1.1. General features of LEVEL5 = 151
5.1.2. Essential Sections in the Knowledge Base = 153
5.1.3. Editing, Compiling, and Running an Application = 156
5.1.4. User Interface in LEVEL5 = 159
5.1.5. User-Interface Development = 161
5.1.6. Treatment of Uncertainty in LEVEL5 = 165
5.1.7. System Control Statements = 167
5.1.8. Outside Hooks in LEVEL5 = 169
5.1.9. Other Features in LEVEL5 = 170
5.2 Programming Languages for Expert Systems = 171
5.2.1. A Brief Review of Prolog(Optional) = 171
5.2.2. A Brief Review of Lisp(Optional) = 177
CHAPTER 6 Inductive Reasoning with 1st-Class = 187
6.1 General Features of 1st-CLASS = 189
6.1.1. Input Requirements for 1st-CLASS = 189
6.1.2. Processing in 1st-CLASS = 190
6.2 Working with 1st-CLASS = 191
6.2.1. First Screen : Files = 191
6.2.2. Second Screen : Definitions = 192
6.2.3. Third Screen : Examples = 196
6.2.4. Fourth Screen : Methods = 197
6.2.5. Fifth Screen : Rule = 198
6.2.6. Sixth Screen : Advisor = 202
6.3 Treatment of Uncertainty in 1st-CLASS(Optional) = 203
6.4 Modular Processing in 1st-CLASS = 205
6.5 Other Features in 1st-CLASS = 208
6.5.1. Methods in 1st-CLASS(Optional) = 208
6.5.2. Outside Hooks(Optional) = 209
6.5.3. Development Tools(Optional) = 210
6.6 Using 1st-CLASS = 211
6.6.1. Inductive Reasoning with 1st-CLASS = 212
6.6.2. Combining 1st-CLASS with Other Methods = 212
CHAPTER 7 System Development and Knowledge Acquisition = 219
7.1 Stages in Developing Expert Systems = 222
7.1.1. System Development Life Cycle = 224
7.1.2. Prototyping = 227
7.2 Systems Analysis in Expert Systems = 229
7.2.1. Problem Definition and Goal Identification = 229
7.2.2. Domain Analysis, Modularization, and Expert Identification = 230
7.2.3. Communication Process = 231
7.3 Knowledge Acquisition as the Logical Design = 233
7.3.1. Logical Design vs. Physical Design of the Knowledge Base = 234
7.3.2. Expert Selection = 234
7.3.3. Sources of Knowledge = 236
7.3.4. Knowledge Acquisition Methods = 236
7.3.5. Knowledge Acquisition Modes = 245
7.3.6. Issues in Multi-expert Knowledge Acquisition = 247
7.3.7. Knowledge Collection Tools = 250
7.3.8. Organizational Aspects of Knowledge Acquisition = 251
7.4 The Physical-Design of Expert Systems = 253
7.4.1. Software Decisions = 253
7.4.2. Hardware Decisions = 255
7.4.3. User-Interface Decisions = 256
7.4.4. The Physical Design of the Knowledge Base = 259
7.5 Coding, Testing, and Reliability of Expert Systems = 260
7.5.1. Managing the Coding Process = 260
7.5.2. Testing = 261
7.5.3. Reliability of Expert Systems = 264
7.6 Implementation and Post-implementation of Expert Systems = 266
7.6.1. Implementation Considerations = 266
7.6.2. Post-implementation Considerations = 267
PART Ⅳ OBJECT-ORIENTED REPRESENTATION AND HYBRID METHODS
CHAPTER 8 Object-Oriented Representation and Design = 277
8.1 The Evolution of Object-Oriented Methods = 280
8.1.1. Semantic Nets = 280
8.1.2. Scripts = 283
8.1.3. Frames = 285
8.2 Object-Oriented Programing(OOP) = 287
8.2.1. The Need for OOP = 288
8.2.2. Class Abstraction = 288
8.2.3. Hierarchy of Classes = 289
8.2.4. Inheritance = 290
8.2.5. Object as an Instance of a Class = 291
8.2.6. Methods = 291
8.2.7. Modularity and Encapsulation = 293
8.2.8. External and Internal Views = 295
8.3 Modeling Knowledge in Objects-based Representation Methods = 298
8.3.1. Object-Oriented Analysis(OOA) = 299
8.3.2. Object-Oriented Design(OOD) = 302
8.4 Logical Design of the Object-Oriented Represention = 302
8.4.1. Designing Classes and Their Relations = 304
8.4.2. Designing Methods(Optional) = 309
8.4.3. Designing the Dynamics of the System = 315
8.4.4. Documentation of the Design = 315
8.4.5. Tools for Object-Oriented Analysis and Design = 315
8.5 Physical Design of the Object-Oriented Representation = 316
8.5.1. Object-Oriented Programming Languages = 317
8.5.2. Conventation vs. Object-Oriented Programming = 318
8.5.3. Categories of OOP Languages = 320
8.5.4. Special Issues in the Physical Design = 324
