Landslide susceptibility mapping using support vector machine and random forest methods in Haraz, Iran

This study evaluates landslide susceptibility in Haraz region of Iran using machine learning methods such as support vector machine (SVM) and random forest (RF). Focusing on vulnerable locations, this research employs a dataset of ۳۰۰۰ points classified using k-fold cross-validation to increase model accuracy. Conditioning factors, including topography, hydrologic, lithology, land cover, and anthropogenic factors, are analyzed to assess their impact on landslide hazard. The RF method with an AUC value of ۰.۹۸ outperformed the SVM method with an AUC of ۰.۹۶, providing a robust framework for landslide susceptibility mapping. The insights of this study aim to support effective risk management and disaster reduction efforts in the region, using advanced machine learning algorithms to predict and manage landslide hazards.

Keywords:

Random Forest , Support Vector Machine , Landslide Risk Mapping , Machine Learning , K-fold cross-validation