A distributed genetic algorithm for graph-based clustering

Abstract

Clustering is one of the most prominent data analysis techniques to structure large datasets and produce a human-understandable overview. In this paper, we focus on the case when the data has many categorical attributes, and thus can not be represented in a faithful way in the Euclidean space. We follow the graph-based paradigm and propose a graph-based genetic algorithm for clustering, the flexibility of which can mainly be attributed to the possibility of using various kernels. As our approach can naturally be parallelized, while implementing and testing it, we distribute the computations over several CPUs. In contrast to the complexity of the problem, that is NP-hard, our experiments show that in case of well clusterable data, our algorithm scales well. We also perform experiments on real medical data. © 2011 Springer-Verlag Berlin Heidelberg.