An expert system is a software package which emulates the decision-making ability of a hu- man expert in a specific domain. For exam- ple, building computer programs for diagnos- ing a patient based on a given set of symptoms and advising him of an appropriate treatment [2], the design and diagnosis of faults in elec- tronic circuits [6], the characterisation of oil and gas fields with an assessment of their hydrocar- bon potential [4] or configuration of computer systems[8]. The underlying knowledge represen- tation scheme for most of these expert systems is production system-sometimes called a rule- based approach. In general, the productions sys- tems are computationally as powerful as other computable functions such as Turing machine, predicate logic, lambda calculus etc. However, it has been argued that if ... then ... kind
of rules are fundamental to capture the human expert, knowledge [7].
It is not difficult to see the commercial values of the expert systems developed during last 20 years. For example, Prospector [ll] was success- fully used in mineral deposit exploration, and XCON [8] has saved DEC from a lot of com- puter configuration work since 1983.
Among many proposed methods for build- ing expert systems, the prototyping approach has been identified as an appropriate procedure. Because the development of expert systems in- volves understanding of human knowledge which can hardly be completed without obtaining feed- back from all relevant parties in the problem en- vironment.
While developing knowledge-based systems, we must address the two fundamental problems: adequately representing and understanding the expert knowledge; and dealing with the dynamic (soft) nature of knowledge. The traditional ap- proaches mainly based on the interaction be- tween the knowledge engineer and the domain expert. However, the role of user in the develop- ment process has been undermined. For exam- ple. how the relationship between user and other agents in the environment affects the process of building human understandable programs that exhibits intelligence. We argue that the human factors involved in the user/expert-system inter- action are fundamental to the development of
An expert system is a software package which emulates the decision-making ability of a hu- man expert in a specific domain. For exam- ple, building computer programs for diagnos- ing a patient based on a given set of symptoms and advising him of an appropriate treatment [2], the design and diagnosis of faults in elec- tronic circuits [6], the characterisation of oil and gas fields with an assessment of their hydrocar- bon potential [4] or configuration of computer systems[8]. The underlying knowledge represen- tation scheme for most of these expert systems is production system-sometimes called a rule- based approach. In general, the productions sys- tems are computationally as powerful as other computable functions such as Turing machine, predicate logic, lambda calculus etc. However, it has been argued that if ... then ... kind of rules are fundamental to capture the human expert, knowledge [7]. It is not difficult to see the commercial values of the expert systems developed during last 20 years. For example, Prospector [ll] was success- fully used in mineral deposit exploration, and XCON [8] has saved DEC from a lot of com- puter configuration work since 1983. Among many proposed methods for build- ing expert systems, the prototyping approach has been identified as an appropriate procedure. Because the development of expert systems in- volves understanding of human knowledge which can hardly be completed without obtaining feed- back from all relevant parties in the problem en- vironment. While developing knowledge-based systems, we must address the two fundamental problems: adequately representing and understanding the expert knowledge; and dealing with the dynamic (soft) nature of knowledge. The traditional ap- proaches mainly based on the interaction be- tween the knowledge engineer and the domain expert. However, the role of user in the develop- ment process has been undermined. For exam- ple. how the relationship between user and other agents in the environment affects the process of building human understandable programs that exhibits intelligence. We argue that the human factors involved in the user/expert-system inter- action are fundamental to the development of
การแปล กรุณารอสักครู่..