Consensus Clustering using Weights of Clusters

Consensus clustering is a technique that leverages an ensemble of clusterings produced by various algorithms to create a single, superior, and robust clustering. In this study, we propose the application of cluster weights to form the consensus, where the weights are derived from cluster-level information gathered from the diversity within the ensemble, as well as data distribution-level information from the internal quality of the clusterings. Our experimental results demonstrate the efficiency of the proposed weighting method compared to other recently proposed consensus clustering algorithms.