Determination of Number of Clusters Using VAT Images and Genetic Algorithms

Determining number of clusters present in a data set is an important problem in clustering. There exist very few techniques that can solve this problem satisfactorily. Some of these techniques rely on user supplied information, while some others use cluster validity indices which are expensive with regard to computation time. This paper proposes an alternative solution for the concerned problem that makes use of the concepts of genetic algorithms, cluster validity indices and a recently developed visual mechanism for determining the clustering tendency (VAT, Visual Assessment of Tendency for clustering). It is shown that the present approach does not require any user supplied information, and is able to find the appropriate number of clusters present in a data set automatically and very efficiently.