Optimizing ensemble learning for satellite-based multi-hazard monitoring and susceptibility assessment of landslides, land subsidence, floods, and wildfires
Analysis reveals optimized Random Forest models improve multi-hazard susceptibility mapping, suggesting significant risk variations across regions.
Key Points
Optimized Random Forest models significantly improve accuracy for assessing landslide and wildfire hazards, indicating enhanced predictive capabilities.
The RF-GA model achieved 91.1% accuracy for floods, while RF-PSO produced 95.9% for land subsidence, demonstrating advanced performance over traditional methods.
Utilizing two meta-heuristic algorithms, Genetic Algorithm and Particle Swarm Optimization, effectively optimizes the Random Forest approach for spatial hazard assessment.
The findings suggest significant interaction between floods and land subsidence, highlighting the need for integrated disaster risk management strategies.
Cite This Study
Razavi-Termeh et al. (2025) studied this question.