Sugeno's fuzzy measure and fuzzy clustering (Q1059349)

Proposed is an iterative clustering method making use of Sugeno's fuzzy measure. The minimized performance index is of the form \(\sum^{c}_{i=1}\sum^{n}_{j=1}g_{ij}d_{ij}\), where c and n stand for the number of clusters (classes) and objects clustered, respectively. \(g_{ij}\) is a fuzzy density attached to the j-th object in the light of the i-th cluster, while \(d_{ij}\) forms a distance function between the object and the relevant centroid of the i-th class.