Exact algorithms for minimum sum-of-squares clustering

Abstract

NP-Hardness of Euclidean sum-of-squares clustering -- Computational complexity -- An incorrect reduction from the K-section problem -- A new proof by reduction from the densest cut problem -- Evaluating a branch-and-bound RLT-based algorithm for minimum sum-of-squares clustering -- Reformulation-Linearization technique for the MSSC -- Branch-and-bound for the MSSC -- An attempt at reproducting computational results -- Breaking symmetry and convex hull inequalities -- A branch-and-cut SDP-based algorithm for minimum sum-of-squares clustering -- Equivalence of MSSC to 0-1 SDP -- A branch-and cut algorithm for the 0-1 SDP formulation -- Computational experiments -- An improved column generation algorithm for minimum sum-of-squares clustering -- Column generation algorithm revisited -- A geometric approach -- Generalization to the Euclidean space -- Computational results.

Uncontrolled Keywords

Optimisation mathématique; Classification automatique (Statistique) -- Mathématique; Moindres carrés