Cluster analysis methods require the users to give the number of clusters in which they would like to group objects. Determination of the actual number of clusters is not easy and proves difficult. In the paper two methods are presented that are helpful in establishing the number of clusters: (1) CCC (Cubic Clustering Criterion) method based on comparison of expected R2 value for data coming from a uniform distribution with the R2 value observed in a sample and (2) the method which is based on quality measures of building clusters in case of fuzzy grouping of data. Both advantages and drawbacks of these methods and examples of their application are presented.
Financed by the National Centre for Research and Development under grant No. SP/I/1/77065/10 by the strategic scientific research and experimental development program:
SYNAT - “Interdisciplinary System for Interactive Scientific and Scientific-Technical Information”.