.png){kind=icon}
.png){kind=icon}
Return
Attention guided spectral feature selection for efficient soil nutrient prediction
.png){kind=icon}
PRE
AI
A
N
W
DOI：10.1016/j.asoc.2026.115310.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Introduces AG-SFS, an end-to-end model for shared-target band selection in soil VNIR–SWIR spectroscopy. • Enhances nutrient prediction accuracy using a systematic band selection approach. • Captures multi-target correlations using Feature-Wise Attention supported by l1 regularization, TV, and entropy penalties. • Achieves computational efficiency by focusing on a subset of bands instead of the entire spectrum.
Keywords:
AG-SFS
band selection
soil nutrient prediction
feature-wise attention
spectral feature selection
Journal
IF:
6.6
Papers:
1.4W
Citations:
4.8W