Lanxensemble: explainability enabled landslide susceptibility prediction

Key Points

• Landslide susceptibility prediction accuracy reached 85% in diverse terrains, indicating high efficiency.
• Key metrics included AUC scores, analysis of outcomes in real-world conditions, and geographical data utilization.
• Observational analysis employs advanced machine learning techniques, specifically ensemble methods for robust predictions.
• Results may guide future tools in disaster management and policy formulation for risk reduction.