MapReduce algorithms for robust center-based clustering in doubling metrics

Enrico Dandolo, Alessio Mazzetto, Andrea Pietracaprina, Geppino Pucci

Abstract: Highlights•First distributed algorithm for center-based clustering with outliers.•Coreset-based approach featuring approximation close to best sequential one.•Scalable algorithm suitable for very large datasets of constant doubling dimension.•Analysis parametric in the intrinsic dimensionality of the input pointset.