Expert Systems Expert Systems Programs which attempt to

Expert Systems

Expert Systems • Programs which attempt to imitate the reasoning processes and knowledge of experts in solving specific problem • Widely used in industry • Apply a direct means of applying expertise • purpose is to make expert knowledge and experience more widely available

Knowledge Engineering • The construction of expert systems is done by using knowledge engineering • This involves – knowledge acquisition - collection – knowledge representation - organisation & representation (IF-THEN rules) – knowledge inferencing – knowledge transfer

Inferencing • making up a new fact based on existing knowledge • Unique feature of ES is ability to reason • expertise is stored in knowledge base • computer is programmed so that it can make inferences • (drawing conclusion from knowledge base) • This is performed in the inference engine • includes procedures regarding problem solving

Rules • Most ES are rule based – Knowledge is stored in the form of rules – problem-solving procedures • IF the engine is idle AND the fuel pressure is less than 38 psi, AND the guage is accurate, THEN there is a fuel system fault

Explanation Capability • Unique ability to explain advice or recommendation • subsystem called the justifier • enables the system to examine its own reasoning and explain its operation

Structure • Development environment – used to build the components and put knowledge into the knowledge base • Consultation environment – used by nonexperts to obtain expert knowledge and advice • Improvement environment

Knowledge Acquisition Subsystem • Accumulation, transfer and transformation of problem-solving expertise from expert sources to a computer • Knowledge engineers gather this expertise • help the expert structure the problem area by interpreting and integrating human answers to – questions – drawing analogies – posing counterexamples

Knowledge Base • Contains the knowledge necessary for – understanding, formulating, and solving problems • Includes facts and rules that use the knowledge to solve specific problems (heuristics) • heuristics express the informal judgmental knowledge in an application area • The information in the knowledge base is incorporated into a computer program by a process called knowledge representation

Inference Engine • Brain of the expert system (rule interpreter) • program provides a methodology for reasoning about information in the knowledge base to formulate conclusions

Explanation Subsystem • Responsibility for conclusions can be traced by answering questions such as: – Why was a certain question asked by the expert system – How was a certain conclusion reached – Why was a certain alternative rejected Knowledge Refining System • Analysis the reasons for success and failure

Development • A tool used to expedite development is called an ES shell • A shell can be used for many applications: insert new knowledge base • Include all the generic components of an ES, but they do not include the knowledge • faster development plus the programming skill required is much lower

Areas of use • Prediction systems – demographic predictions, economic forecasting • Diagnostic systems – medical, software diagnosis • Monitoring systems – compare observations of system behaviour with standards • Control systems – govern the overall behaviour of a system

Reference Expert System applications in business: A review and analysis of the literature Bo K. Wong, John A. Monaco Information and Management Vol 29 Part 3 1995 HF 5548. 125. 15

Benefits • • Increased output and productivity Accessibility to knowledge Decreased decision-making time Reduced downtime Capture of Scarce expertise Flexibility Elimination of the need for expensive equipment • Operations in hazardous environments

Benefits • Increased Capabilities of other systems • Integration of several experts opinions • Ability to work with incomplete or uncertain information • Provide training • Improved decision making • Transfer knowledge to remote locations

Limitations • Knowledge is not always readily available • Expertise can be hard to extract from humans • The approach of each expert to situation assessment may be different, yet correct • It is hard, even for a highly skilled expert, to abstract good assessments under pressure • ES work well only in a narrow domain of knowledge

Limitations • Vocabulary, jargon that experts use for expressing facts and relations is often limited and not understood by others • Help is required from knowledge engineers who are rare and expensive • lack of trust by end users