New validity index for determining the number of clusters in K-means clustering
LI Shuang-hu, Wang Tie-hong · Hebei Sheng Kexueyuan xuebao · 2003
K-Means Clustering Algorithm is one of the most popular methods in cluster analysis. However, it is well known that K-means algorithm suffers from initial starting conditions effects(initial clustering and instance order effects). For more detailed discussion on initialization methods, see literature \. Another weakness of k-means algorithm is that the number of clusters, k, must be supplied as a parameter. In this paper, a new validity measure for k-means clustering is presented to allow the number of clusters to be determined automatically.