Fuzzy Dissimilarity Learning in Case-Based Reasoning

Case-based reasoning (CBR) attempts to solve new problems by using previous experiences. However traditional CBR systems are restricted by the similarity requirement, i.e., the availability of similar cases to new problems. This paper proposes a novel CBR approach that exploits dissimilarity information in problem solving. A fuzzy dissimilarity model consisting of fuzzy rules has been developed for assessing dissimilarity between cases. Further, it is indicated that the construction of fuzzy dissimilarity rules can be realized by learning from the case library. Empirical studies have demonstrated that fuzzy dissimilarity models can be built upon a small case library while still yielding competent performance of the CBR system. Key-Words: case-based reasoning; similarity; dissimilarity; case library; fuzzy rles; fuzzy reasoning