Scholarlink

<p>This research aimed to investigate the effectiveness of using physics-based metaheuristic algorithms in combination with ensemble machine-learning models for landslide susceptibility mapping (LSM). By optimizing two ensemble machine learning models (Random Forest (RF) and eXtreme Gradient Boosting (XGBoost)) using three physically based metaheuristic algorithms (Archimedes Optimization Algorithm (ArchOA), Chernobyl Disaster Optimizer (CDO), and Simulated Annealing (SA)), the study demonstrated the potential of this approach to improve the accuracy of landslide susceptibility maps significantly. This research employed an Explainable Artificial Intelligence (XAI) method, namely SHAP (SHapley Additive explanations), to interpret developed models. Additionally, an uncertainty quantification approach was utilized to assess the reliability of susceptibility maps. The landslide susceptibility map was prepared using 153 landslide occurrence points and 15 effective criteria on landslides in Khalkhal town of Iran. The results revealed that XGBoost-CDO, XGBoost-ArchOA, RF-CDO, XGBoost-SA, XGBoost, RF-SA, and RF models, respectively, were the most accurate, with the area under the curve (AUC) of the Receiver Operating Characteristic (ROC) values of 0.952, 0.958, 0.951, 0.938, 0.929, 0.847, and 0.842. The results of the uncertainty map showed that areas with higher landslide susceptibility have lower uncertainty, with 65.35% of all landslides occurring in regions characterized by high landslide susceptibility and lower uncertainty.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Investigating the efficacy of physics-based metaheuristic algorithms in combination with explainable ensemble machine-learning models for landslide susceptibility mapping

  • 期刊论文

  • 发表日期：

    2025年2月8日

查看全文
  • Seyed Vahid Razavi-Termeh,
  • Abolghasem Sadeghi-Niaraki,
  • Rizwan Ali Naqvi,
  • Soo-Mi Choi

摘要

This research aimed to investigate the effectiveness of using physics-based metaheuristic algorithms in combination with ensemble machine-learning models for landslide susceptibility mapping (LSM). By optimizing two ensemble machine learning models (Random Forest (RF) and eXtreme Gradient Boosting (XGBoost)) using three physically based metaheuristic algorithms (Archimedes Optimization Algorithm (ArchOA), Chernobyl Disaster Optimizer (CDO), and Simulated Annealing (SA)), the study demonstrated the potential of this approach to improve the accuracy of landslide susceptibility maps significantly. This research employed an Explainable Artificial Intelligence (XAI) method, namely SHAP (SHapley Additive explanations), to interpret developed models. Additionally, an uncertainty quantification approach was utilized to assess the reliability of susceptibility maps. The landslide susceptibility map was prepared using 153 landslide occurrence points and 15 effective criteria on landslides in Khalkhal town of Iran. The results revealed that XGBoost-CDO, XGBoost-ArchOA, RF-CDO, XGBoost-SA, XGBoost, RF-SA, and RF models, respectively, were the most accurate, with the area under the curve (AUC) of the Receiver Operating Characteristic (ROC) values of 0.952, 0.958, 0.951, 0.938, 0.929, 0.847, and 0.842. The results of the uncertainty map showed that areas with higher landslide susceptibility have lower uncertainty, with 65.35% of all landslides occurring in regions characterized by high landslide susceptibility and lower uncertainty.