A genetic algorithm for cluster analysis

Evolving clusters in gene-expression data

“…Section 2 formalizes the clustering problem to be tackled by means of evolutionary algorithms. Section 3 reviews the clustering genetic algorithm (CGA) [23], whereas Section 4 proposes improvements on it. Section 5 reports results involving six gene-expression datasets.…”

“…The clustering genetic algorithm (CGA) [23], originally designed to optimize both the number of clusters and the corresponding data partition, is adopted here as a starting point algorithm, from which more efficient evolutionary algorithms for clustering gene-expression data are derived. These algorithms were briefly introduced in a previous work [22], in which five bioinformatics datasets were used for performance assessment.…”

A genetic k-medoids clustering algorithm

“…GAs are search procedures that use the mechanics of evolution and natural genetics. They have been earlier reported for the clustering problem (Lucasius et al, 1993;Murthy and Chowdhury, 1996;Scheunders, 1997;Estivill-Castro and Murray, 1997;Falkenauer, 1998;Hall et al, 1999;Krishna, 1999;Maulik and Bandyopadhyay, 2000;Bandyopadhyay and Maulik, 2000;Hruschka and Nelson, 2003;Yi, 2004). However, they may take a large amount of time to converge and/or assume the number of clusters to be fixed a priori.…”

On the efficiency of evolutionary fuzzy clustering

“…Standard genetic operators are usually not suitable for clustering problems for several reasons, as detailed in Falkenauer (1998), Hruschka and Ebecken (2003), Hruschka et al (2006). In brief, such operators often just manipulate objects by means of their corresponding cluster labels, without taking into account their connections with other clusters.…”