In this paper, a multi-objective clustering technique is proposed to find the appropriate partition of a set of objects directly from dissimilarity data as well as automatically determine the proper number of clusters. We propose a clustering technique based on the explicit optimization of a partitioning with respect to multiple, complementary clustering objectives which take into account both heterogeneity for each class and isolation between classes, as described by Vichi (2008). The algorithm will automatically attain a set of Pareto-optimal solutions in terms both of the number of clusters and the appropriate classification of objects into groups.
In this paper, a multi-objective clustering technique is proposed to find the appropriate partition of a set of objects directly from dissimilarity data as well as automatically determine the proper number of clusters. We propose a clustering technique based on the explicit optimization of a partitioning with respect to multiple, complementary clustering objectives which take into account both heterogeneity for each class and isolation between classes, as described by Vichi (2008). The algorithm will automatically attain a set of Pareto-optimal solutions in terms both of the number of clusters and the appropriate classification of objects into groups.
Multi-objective genetic algorithm based clustering for dissimilarity data / Bocci, Laura. - STAMPA. - (2010), pp. 271-272. (Intervento presentato al convegno Joint meeting GfKl-CLADAG 2010 tenutosi a Firenze (Italy) nel September 8-10, 2010).
Multi-objective genetic algorithm based clustering for dissimilarity data
BOCCI, Laura
2010
Abstract
In this paper, a multi-objective clustering technique is proposed to find the appropriate partition of a set of objects directly from dissimilarity data as well as automatically determine the proper number of clusters. We propose a clustering technique based on the explicit optimization of a partitioning with respect to multiple, complementary clustering objectives which take into account both heterogeneity for each class and isolation between classes, as described by Vichi (2008). The algorithm will automatically attain a set of Pareto-optimal solutions in terms both of the number of clusters and the appropriate classification of objects into groups.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
