The article analyses the flaws in the tranditional CBR system firstly, and brings forward a new approach and arithmetic of CBR system. That is dividing the case-base into urser-base and application-base, every time when search the case-base, we can search the corresponding case-base according the users, which can faster speed of searching.
The basic and difficult problem is the definement or measurement of the similitude between two cases. The tranditional CBR system always adopts the...