Seminář: Case–Based Reasoning in the NEST System

The paper describes the basic form of case–based reasoning as it was implemented in the expert system NEST (originally a rule–based system). The reasoning proceeds over the cases storage, which (contrary to the rule knowledge base) needn’t be created by expert. There are two types of case–based reasoning in the system – compositional and logical. These types differ in the way of using the cases storage, and inspired by the compositional and logical approach to uncertainty processing in the original NEST system. Eventually, future extensions of the system will be discussed, especially, the integration of rule-based and case-based reasoning and the considerations on major overhaul of the system.