Consensus Clustering Using Weight of Clusters and Clusterings: A Dual-Weighted Approach

This paper presents a novel consensus clustering framework that integrates both cluster-level and clustering-level weighting strategies. Traditional consensus clustering methods either weight the clusters or the base clusterings, but often fail to optimally combine these two strategies. We propose a dual-weighting scheme where weights are assigned to clusters based on internal and external consistency, and to the base clusterings based on their agreement with the ensemble. By applying a combined weight, we ensure that both high-quality clusters and consistent clusterings contribute more to the final consensus. Experimental results on several benchmark datasets demonstrate the superiority of the proposed method over existing clustering ensemble techniques.