CommonKADS: Course Syllabus

PART I: Introducing CommonKADS

INTRODUCTION TO KNOWLEDGE ENGINEERING: Why are KBS different? Knowledge engineering methods; the CommonKADS methodology.

THE EXPERTISE MODEL: Components of the CommonKADS Expertise Model; rationale for the model; modelling languages.

OTHER COMMONKADS MODELS: Organisational model; task, agent and communication models; design model.

INTRODUCTION TO ILOG KADSTOOL: Facilities available in KADSTool; how to use KADSTool.

PART II: Feasibility

SELECTING AND SCOPING A KBS APPLICATION: Applying the CommonKADS organisational and task models; an expanded checklist for identifying feasibility.

PART III: Knowledge Elicitation

INTRODUCTION TO KNOWLEDGE ELICITATION: What is knowledge elicitation; why is it difficult; declarative vs. procedural knowledge; characterising different types of expert.

THE INTERVIEW: Knowledge elicitation through interviews; suggested approach; transcript analysis.

ACQUIRING PROCEDURAL KNOWLEDGE: Acquiring and analysing a protocol; using the ``20 Questions'' technique.

LADDERED GRIDS: Using the ``laddered grid'' technique.

MULTI-DIMENSIONAL TECHNIQUES: The card sorting technique; the repertory grid technique;

PART IV: CommonKADS Expertise Modelling

DOMAIN KNOWLEDGE: CommonKADS domain ontology; domain models; Conceptual Modelling Language; abstraction to the ``model level''.

INFERENCE KNOWLEDGE: Inference structures; library of generic inference structures; configuring and instantiating generic inference structures.

TASK KNOWLEDGE: Task structures; Problem Solving knowledge.

APPROACHES TO DEVELOPING THE EXPERTISE MODEL: Project Management Activity Cycle; four approaches to developing an expertise model.

PART V: Design and Implementation

KBS DESIGN: CommonKADS Design modelling; different ways of decomposing an expertise model; using ``probing questions'' for architectural design.

VALIDATION, VERIFICATION AND MAINTENANCE OF A KBS: Performing validation and verification: a) during knowledge acquisition, b) on a CommonKADS Expertise Model and c) in rule-based systems.