FPF-SB: A SCALABLE ALGORITHM FOR MICROARRAY GENE EXPRESSION DATA CLUSTERING

Efficient and effective analysis of large datasets from microarraygene expression data is one of the keys to time-critical personalizedmedicine. The issue we address here is the scalability of the dataprocessing software for clustering gene expression data into groupswith homogeneous expression profile. In this paper we propose FPFSB,a novel clustering algorithm based on a combination of theFurthest-Point-First (FPF) heuristic for solving the k-center problemand a stability-based method for determining the number of clustersk. Our algorithm improves the state of the art: it is scalable to largedatasets without sacrificing output quality.