论文部分内容阅读
A new method (the Contrast statistic) for estimating the number of clusters in a set of data is proposed. The technique uses the output of self-organising map clustering algorithm, comparing the change in dependency of “Contrast” value upon clusters numbe