Determinação do número de agrupamentos em conjuntos de dados multidimensionais utilizando algoritmos genéricos

Abstract

The cluster analysis has been used successfully in several research areas with the objective of grouping similar data according to their features. An efficient technique in cluster analysis is the use of the probability density function that presents the number of clusters graphically. The genetic algorithms were used with success to inform the number of clusters in a group of unidimensional data. The existent methods for the cluster analysis in multidimensional data need an approximate number of clusters to locate them. The acting of these methods depends directly on this number of clusters. The purpose of this work is to use the genetic algorithms to predict the number of clusters in multidimensional data.