3Center of Soft Computing and Intelligent System Studies,

Various concepts of similarity relation have been already proposed dealing with number of coin-cided attributes, Euclid distance, etc. In 2003, we proposed a concept of similarity by applying a uniqueness measure, called uniqueness-based similarity. The concept of similarity is based on human’s perception dealing with unique value of attributes given a subset of objects in an infor-mation system. The subset of objects is regarded as knowledge by which it is possible to have a different degree of similarity between two given objects. For example, in a certain knowledge, object A is more similar to object B than object C. However, in the other knowledge, object A may be more similar to object C than object B. Based on uniqueness-based similarity, this paper introduces a concept of clustering and examines its properties