DocumentCode :
2668397
Title :
On evaluation of clustering using homogeneity analysis
Author :
Sato-Ilic, Mika
Author_Institution :
Tsukuba Univ., Ibaraki, Japan
Volume :
5
fYear :
2000
fDate :
2000
Firstpage :
3588
Abstract :
Proposes a new technique to evaluate the results of fuzzy clustering. Fuzzy clustering is a method to obtain “natural groups” from given observations by using the assumption of a fuzzy subset on clusters. By this method, we can obtain the “degree of belongingness” of an object to a cluster, i.e. each object can belong to several clusters with several degrees, and the boundaries of the clusters become uncertain. However, such clustering sometimes causes confusion when trying to understand the clustering behavior, because the clusters are not exclusive, so each cluster has its own property of mixing with other clusters. Sometimes, a very mixed cluster is created, which makes it difficult to interpret the results. In this case, we need to evaluate the cluster homogeneity in the sense of the degree of belongingness. So, in this paper, I propose an evaluation technique for the results of fuzzy clustering using homogeneity analysis. Several numerical examples demonstrate the validity of this method
Keywords :
fuzzy set theory; pattern clustering; belongingness degree; cluster homogeneity analysis; cluster mixing property; evaluation technique; fuzzy clustering; fuzzy subset; mixed clusters; natural groups; uncertain clusters; Uncertainty;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Systems, Man, and Cybernetics, 2000 IEEE International Conference on
Conference_Location :
Nashville, TN
ISSN :
1062-922X
Print_ISBN :
0-7803-6583-6
Type :
conf
DOI :
10.1109/ICSMC.2000.886566
Filename :
886566
Link To Document :
بازگشت