Density-Sensitive Evolutionary Clustering

In this study, we propose a novel evolutionary algorithm-based clustering method, named density-sensitive evolutionary clustering (DSEC). In DSEC, each individual is a sequence of real integer numbers representing the cluster representatives, and each data item is assigned to a cluster representative according to a novel density-sensitive dissimilarity measure which can measure the geodesic distance along the manifold. DSEC searches the optimal cluster representatives from a combinatorial optimization viewpoint using evolutionary algorithm. The experimental results on seven artificial data sets with different manifold structure show that the novel density-sensitive evolutionary clustering algorithm has the ability to identify complex non-convex clusters compared with the K-Means algorithm, a genetic algorithm-based clustering, and a modified K-Means algorithm with the density-sensitive distance metric.

• Type: Conference
• Year: 2007
• Where: PAKDD
• Authors: Maoguo Gong, Licheng Jiao, Ling Wang, Liefeng Bo
• Keywords: Cluster Representatives, Data Mining, Density-sensitive Evolutionary Clustering, Evolutionary Algorithm-based Clustering, PAKDD 2007