Knowledge Induction Based on Randomization in Case-Based Reasoning
Thouraya Bouabana‐Tebibel, Stuart H. Rubin, Asmaa Chebba, Sofiane Bediar, Syphax Iskounen · 2016
Case-Based Reasoning (CBR) interests the scientific community, whom are concerned with scalability in knowledge representation and processing. CBR systems scale far better than rule-based systems. Rule-based systems are limited by the need to know the rules of engagement, which is practically unobtainable. The work presented in this paper pertains to knowledge generalization based on randomization. Inductive knowledge is inferred through subsumption and transmutation rules. Knowledge is dynamically generated, thus allowing for a gain in the inferential space and research time. It is validated based on the domain user expertise. The approach is illustrated with an example and is seen to be properly implemented and tested.