Partial multi-label feature selection via adaptive dual-graph regularization

doi:10.1016/j.knosys.2025.114077

Partial multi-label feature selection via adaptive dual-graph regularization

2025-07-10

0

PRE

AI

DOI：10.1016/j.knosys.2025.114077

Original

Original request for help

Save

Abstract

En 中文

• A novel method called Partial Multi-label Feature Selection via Adaptive Dual-graph Regularization is proposed, embedding label disambiguation and the selection of a discriminative feature subset into the same framework. • An adaptive dual-graph regularization is introduced to simultaneously explore the non-linear geometric information of the ground-truth label space and the feature space, thereby more accurately recovering the ground-truth label distribution. • An effective optimization algorithm is designed to solve the optimization problem of PMFS-ADG, ensuring convergence. • A comprehensive set of experiments on diverse synthetic and genuine partial multi-label datasets is carried out to highlight the robustness and effectiveness of the proposed method.

Keywords:

Feature selection

Partial multi-label learning

Adaptive dual-graph regularization

Noisy labels

Partial multi-label feature selection via adaptive dual-graph regularization

Abstract

Journal

Organization