Fuzzy Similarity-based Models in Case-based Reasoning
Publication Type:Conference Paper
Source:The 2002 IEEE World Congress on Computation Intelligence, Volume Proceeding, p.1348-1353 (2002)
This paper deals with fuzzy similarity-based models of the basic principle of case-based reasoning (CBR) stating that "similar problems lead (or may lead) to similar outcomes". A stronger form of this principle stating that "outcome attributes are at least as similar as problem description attributes" has been studied in some previous works. In this paper another form of the basic principle stating that "the more similar are the problem description attributes, the more similar are the outcome attributes" is studied. These two forms of the CBR principle are used to infer possible outcomes for the current problem from the information stored in the memory of precedent cases.