The probabilistic method of finding the local-optimum of clustering

The stability of clustering methods is a commonly used approach in cluster analysis for determining the “true” number of groupings. The acceptable clustering is such data sample grouping that is robust to random perturbations of investigated data. In this paper, we propose an algorithm for determining the number of clusters based on the introduction of the initial dataset which are expanded by adding the set of perturbated initial dataset. Refs 30. Figs 2.