Scalable unsupervised feature selection via weight stability

ArXi:2506.06114v4 Announce Type: replace Unsupervised feature selection is critical for improving clustering performance in high-dimensional data, where irrelevant features can obscure meaningful structure. In this work, we