Seismic Site Effects and Aptitude to Urbanization for the Study Area of Taroudant, South-West of Morocco DOI
Sliman Hitouri,

El Arbi Toto,

Mohamed Hafid

и другие.

Опубликована: Янв. 1, 2024

Язык: Английский

Quantifying the Geomorphological Susceptibility of the Piping Erosion in Loess Using LiDAR-Derived DEM and Machine Learning Methods DOI Creative Commons
Sisi Li, Sheng Hu, Lin Wang

и другие.

Remote Sensing, Год журнала: 2024, Номер 16(22), С. 4203 - 4203

Опубликована: Ноя. 11, 2024

Soil piping erosion is an underground soil process that significantly underestimated or overlooked. It can lead to intense and trigger surface processes such as landslides, collapses, channel erosion. Conducting susceptibility mapping a vital way identify the potential for erosion, which of enormous significance water conservation well geological disaster prevention. This study utilized airborne radar drones survey map 1194 sinkholes in Sunjiacha basin, Huining County, on Loess Plateau Northwest China. We identified seventeen key hydrogeomorphological factors influence sinkhole used six machine learning models—support vector (SVM), logistic regression (LR), Convolutional Neural Network (CNN), K-Nearest Neighbors (KNN), random forest (RF), gradient boosting decision tree (GBDT)—for assessment loess sinkholes. then evaluated validated prediction results various models using area under curve (AUC) Receiver Operating Characteristic Curve (ROC). The showed all these algorithms had AUC more than 0.85. GBDT model best predictive accuracy (AUC = 0.94) migration performance 0.93), it could find with high very levels areas. suggests suited fine-scale regions.

Язык: Английский

Процитировано

0

Performance Assessment of Individual and Ensemble Learning Models for Gully Erosion Susceptibility Mapping in a Mountainous and Semi-Arid Region DOI Creative Commons

Meryem El Bouzekraoui,

Abdenbi Elaloui, Samira Krimissa

и другие.

Land, Год журнала: 2024, Номер 13(12), С. 2110 - 2110

Опубликована: Дек. 6, 2024

High-accuracy gully erosion susceptibility maps play a crucial role in vulnerability assessment and risk management. The principal purpose of the present research is to evaluate predictive power individual machine learning models such as random forest (RF), decision tree (DT), support vector (SVM), ensemble approaches stacking, voting, bagging, boosting with k-fold cross validation resampling techniques for modeling Oued El Abid watershed Moroccan High Atlas. A dataset comprising 200 points, identified through field observations high-resolution Google Earth imagery, was used, alongside 21 conditioning factors selected based on their importance, information gain, multi-collinearity analysis. exploratory results indicate that all derived had good accuracy both models. Based receiver operating characteristic (ROC), RF SVM better performances, AUC = 0.82, than DT model. However, significantly outperformed Among ensembles, RF-DT-SVM stacking model achieved highest accuracy, an value 0.86, highlighting its robustness superior capability. prioritization also confirmed best. These findings highlight superiority over ones underscore potential application similar geo-environmental contexts.

Язык: Английский

Процитировано

0

Seismic Site Effects and Aptitude to Urbanization for the Study Area of Taroudant, South-West of Morocco DOI
Sliman Hitouri,

El Arbi Toto,

Mohamed Hafid

и другие.

Опубликована: Янв. 1, 2024

Язык: Английский

Процитировано

0