Title :
A method for determining the number of clusters based on graph theory
Author_Institution :
Sch. of Mech. Eng., Hunan Univ. Of Int. Econ., Changsha, China
Abstract :
This paper mainly addresses the issue of determining the number of clusters. Over a data set, a minimum spanning tree is constructed, and then m-ary tree is employed to search clusters on the minimum spanning tree. The number of large and well separate m-ary tree is the number of clusters. Numerical experiments show that the performance of the proposed method is promising.
Keywords :
graph theory; pattern clustering; trees (mathematics); cluster validity index; graph theory; integrated completed likelihood criterion; m-ary tree; minimum spanning tree; Clustering algorithms; Indexes; Optimization; Pattern recognition; Pediatrics; Probabilistic logic; Shape; cluster validity index; integrated completed likelihood criterion (ICL); m-ary tree; minimum spanning tree;
Conference_Titel :
Fuzzy Systems and Knowledge Discovery (FSKD), 2010 Seventh International Conference on
Conference_Location :
Yantai, Shandong
Print_ISBN :
978-1-4244-5931-5
DOI :
10.1109/FSKD.2010.5569610
