Knowledge Acquisition in CBR for Manufacturing

摘要

This paper describes knowledge acquisition in InfoFrax~(TM) an application of CBR System, in a factory, which manufactures fused cast (electro cast) refractory blocks used as lining for glass melting furnace tanks. Each block is an instance of a case in the case base. The attributes, which are associated with the manufacturing process, are the features of the case. The domain experts cannot emphatically state the relative importance of the features. We propose a method of using CBR to learn the importance of the attributes by attaching weights to the attributes and learning from the changes carried out by the domain experts. The domain experts have accepted this as a solution to the problem of knowledge acquisition and supported developing a system, which will reason like a domain expert.