Aggregation-based self-supervised evidential clustering for imbalanced data

doi:10.1016/j.inffus.2025.103721

Aggregation-based self-supervised evidential clustering for imbalanced data

2025-09-11

0

PRE

AI

DOI：10.1016/j.inffus.2025.103721

Original

Original request for help

Save

Abstract

En 中文

• Proposes a density-based aggregation rule to fuse sub-clusters into singleton clusters until matching true cluster counts. • Introduces self-supervised evidential partitioning via KNN-weighted means for remaining objects using singleton clusters as prior knowledge. • Models partial imprecision in imbalanced data through meta-clusters, reducing error rates; validated on synthetic, real-world, and image datasets.

Aggregation-based self-supervised evidential clustering for imbalanced data

Abstract

Journal

Organization

Cited Papers