Geological Disaster Susceptibility Evaluation of a Random-Forest-Weighted Deterministic Coefficient Model

An assessment of regional vulnerability to geological disasters can directly indicate the extent and intensity risks within study area; thus, providing precise guidance for disaster management efforts. However, in evaluation susceptibility using a single deterministic coefficient model, direct superimposition values each factor, without considering their objective weights, impact accuracy zoning outcomes. To address this limitation, research proposes novel approach: random-forest-weighted model. In method, weight factor is calculated based on model parameter-optimized random forest By weighting superimposing comprehensive map generated. This further divided natural breakpoint method obtain map. validate results, partition statistics ROC (Receiver Operating Characteristic) curve test sample points are utilized. The findings demonstrate that performs well evaluating Huize County. results considered reliable accurate, highlighting effectiveness proposed approach assessing region.