.png){kind=icon}
.png){kind=icon}
Return
Label importance-driven multi-granularity framework for multi-label feature selection
.png){kind=icon}
PRE
AI
Y
C
M
S
Z
B
DOI：10.1016/j.eswa.2026.133918.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• A label importance measure based on information concentration is proposed. • WPS and NPC jointly characterize label-combination similarities. • Multi-level granulation enables discriminative feature evaluation. • Experiments on 14 datasets demonstrate superior performance.
Keywords:
Multi-label
Feature selection
Label importance
Multi-granularity
Journal
IF:
7.5
Papers:
2.9W
Citations:
10.2W
Organization
Cited Papers