Abstract
Case Base Maintenance (CBM) presents one of the key factors success for Case Based Reasoning (CBR) systems. Thence, several CBM policies are proposed to improve their problem-solving performance and competence. However, to the best of our knowledge, all of them are not able to make use of prior knowledge which can be offered by domain experts, especially that CBR is widely applied in real-life domains. For instance, given symptoms of two different cases in medicine area, the doctor can affirm that these two cases should never follow the same treatment, or conversely. This kind of prior knowledge is presented in form of Cannot-Link and Must-link constraints. In addition, most of them cannot manage uncertainty in cases during CBM. To overcome this shortcoming, we propose, in this paper, a CBM policy that handles constraints to exploit experts’ knowledge during case base learning along with managing uncertainty using the belief function theory. This new CBM approach consists mainly in noisy and redundant cases deletion.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
