Title :
Statistical cluster validity indexes to consider cohesion and separation
Author :
Kyung Mi Lee ; Keon Myung Lee ; Chan Hee Lee
Author_Institution :
Dept. of Comput. Sci., Chungbuk Nat. Univ., Cheongju, South Korea
Abstract :
Clusters for a collection of data could be constructed in various shapes according to the employed clustering method and its parameters. To evaluate the quality of clustering results, a variety of cluster validity indexes have been proposed. This paper presents a new cluster validity index which statistically considers both cohesion and separation of clusters. It presents its properties and shows some experimental results.
Keywords :
pattern clustering; statistical analysis; cluster cohesion; cluster separation; clustering method; clustering parameter; data collection; statistical cluster validity index; Clustering algorithms; Clustering methods; Data mining; Data models; Educational institutions; Indexes; Partitioning algorithms; clustering; cohesion; data analysis; separation; validity index;
Conference_Titel :
Fuzzy Theory and it's Applications (iFUZZY), 2012 International Conference on
Conference_Location :
Taichung
Print_ISBN :
978-1-4673-2057-3
DOI :
10.1109/iFUZZY.2012.6409706
