VALIDITY MEASURES FOR HEURISTIC POSSIBILISTIC CLUSTERING

Authors

  • Dmitri A. Viattchenin United Institute of Informatics Problems of the National Academy of Sciences of Belarus

Keywords:

clustering, fuzzy cluster, membership degree, allotment, index of fuzziness, density of fuzzy cluster, cluster validity.

Abstract

A direct algorithm of possibilistic clustering is the effective tool for the data analysis. The approach is based on the concept of allotment among fuzzy clusters. To establish the number of clusters in a data set, some validity measures are presented in this paper. Illustrative examples of application of proposed validity measures to some wellknown data sets are given in comparison with a well-known cluster validity index for objective function-based fuzzy clustering algorithms. Preliminary conclusions are formulated.

Downloads

Published

2010-12-16

Issue

Vol. 39 No. 4 (2010)

Section

Articles

License

Copyright terms are indicated in the Republic of Lithuania Law on Copyright and Related Rights, Articles 4-37.