Learning a Robust Consensus Matrix for Clustering Ensemble via Kullback-Leibler Divergence Minimization

Clustering ensemble has emerged as an important extension of the classical clustering problem. It provides a framework for combining multiple base clusterings of a data set to generate a final consen-sus result. Most existing clustering methods sim-ply combine clustering results without taking into account the noises, which may degrade the cluster-ing performance. In this paper, we propose a novel robust clustering ensemble method. To improve the robustness, we capture the sparse and symmetric er-rors and integrate them into our robust and consen-sus framework to learn a low-rank matrix. Since the optimization of the objective function is diffi-cult to solve, we develop a block coordinate descent algorithm which is theoretically guaranteed to con-verge. Experimental results on real world data sets demonstrate the effectiveness of our method.