A Heuristic Possibilistic Approach to Clustering for Asymmetric Data

This paper deals with the problem of clustering of asymmetric data. A method of the problem solving is based on the application of a direct possibilistic clustering algorithm based on the concept of allotment among fuzzy cluster to a matrix of fuzzy tolerance, which correspond to the set of objects, for which asymmetric distances or proximities hold. The paper provides the description of the method of asymmetric data preprocessing for construction of a matrix of fuzzy tolerance and basic ideas of the method of clustering. An illustrative example of asymmetric data preprocessing and clustering is given and an analysis of the experimental results of the method's application to the Sato-Ilic and Jain's asymmetric data is carried out. Preliminary conclusions are discussed.