.png){kind=icon}
.png){kind=icon}
Return
A machine learning hail recognition method integrating DBSCAN clustering and spatiotemporal tracking
.png){kind=icon}
PRE
AI
Y
C
B
H
S
J
R
H
DOI：10.1016/j.atmosres.2026.109069.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
• Machine-Learning-Based Hail Recognition with Enhanced Accuracy • A novel lightning-centric hail identification framework is proposed, integrating enhanced DBSCAN clustering and spatiotemporal tracking to convert discrete lightning data into coherent storm entities without relying on radar input. • Dynamic lightning parameters and static parameter show statistically significant differences between hailstorms and ordinary thunderstorms. • The LightGBM model trained on six key features achieves excellent performance, with dynamic parameters identified as primary contributors via SHAP analysis.
Keywords:
hail recognition
DBSCAN clustering
spatiotemporal tracking
LightGBM
lightning parameters
Journal
IF:
4.4
Papers:
1.1K
Citations:
2.2W
Organization
Cited Papers