deep-clustering

This paper investigates deep clustering methods on imbalanced tabular data and proposes two novel ensemble approaches that aggregate clustering assignments across embedding dimensions or via majority voting, outperforming individual methods on 16 datasets.