Monte Carlo Positive-Unlabeled Two-Step Learning for Landslide Susceptibility Mapping under Inventory Uncertainty: Comparison with Hard-Threshold PU Two-Step and EIV in Northern Morocco
This study introduces MC-PU-2S, a Monte Carlo-based Positive-Unlabeled two-step learning framework that outperforms deterministic EIV and hard-threshold PU methods in landslide susceptibility mapping for Northern Morocco by effectively propagating inventory uncertainty to generate better-calibrated predictions and essential uncertainty metrics.