A high-accuracy stacking model for automatic identification of aeolian saltating tracks in high-speed pictures

2026-01-09

PRE

Mengjie Pu

Fanmin MEI *

Chuan Lin

Mingjun Sui

Hongda Chen

Shenyang He

Weikai Wang

Chang Zhou

Jin Su

J. Chen

DOI：10.1016/j.aeolia.2026.101031

Original

Original request for help

Save

Abstract

En 中文

• A representative dataset with >10,000 saltation tracks, 17 features, and balanced false/true labels using SMOTE. • A high-accuracy stacking model identifies saltating tracks using CatBoost, Extra Trees, LightGBM, XGBoost, and RF. • Key variables differ between base learners and the meta-model, revealing distinct machine-learning mechanisms.

A high-accuracy stacking model for automatic identification of aeolian saltating tracks in high-speed pictures

Abstract

Journal

Organization

Cited Papers

Citing Papers