2.3 Knowledge Representation and Utilisation
In addition to knowledge acquisition, there are two more key issues we have to consider while building an expert system: knowledge representation and utilisation of the acquired knowledge. A number knowledge representa- tion schemes available to the knowledge engi- neer such as production rules, semantic network, frames, logic, and objected-oriented extensions of existing knowledge representation languages. Knowledge utilisation can also be viewed as knowledge maintenance. In another words, it re- quires an up to date knowledge base which can be used sensibly and consistently. The knowledge that human experts possess is dynamic. Normally, expert systems are built incrementally in the sense that knowledge can be added and revised from time to time. In conventional problem solving approaches such as operational research techniques, knowledge is embedded in the algorithm that may deal with the problem with a specific problem environ- ment. Changes in the knowledge used are dif- ficult to make and usually result in re-designing the whole algorithm. On the contrary, the rep- resentation of knowledge is usually separated clearly from the inference mechanism or reason-
81
ing system in an expert system, making it more convenient to add or remove knowledge without altering the inference mechanism.
2.3 Knowledge Representation and Utilisation In addition to knowledge acquisition, there are two more key issues we have to consider while building an expert system: knowledge representation and utilisation of the acquired knowledge. A number knowledge representa- tion schemes available to the knowledge engi- neer such as production rules, semantic network, frames, logic, and objected-oriented extensions of existing knowledge representation languages. Knowledge utilisation can also be viewed as knowledge maintenance. In another words, it re- quires an up to date knowledge base which can be used sensibly and consistently. The knowledge that human experts possess is dynamic. Normally, expert systems are built incrementally in the sense that knowledge can be added and revised from time to time. In conventional problem solving approaches such as operational research techniques, knowledge is embedded in the algorithm that may deal with the problem with a specific problem environ- ment. Changes in the knowledge used are dif- ficult to make and usually result in re-designing the whole algorithm. On the contrary, the rep- resentation of knowledge is usually separated clearly from the inference mechanism or reason- 81 ing system in an expert system, making it more convenient to add or remove knowledge without altering the inference mechanism.
การแปล กรุณารอสักครู่..
