Code and dataset for: Interpretable machine learning framework for urban flood susceptibility assessment in Yancheng

This repository contains the code and dataset for the paper "Interpretable machine learning framework for urban flood susceptibility assessment: A multi-model comparison with spatial heterogeneity analysis in Yancheng" (Scientific Reports). Contents:- train_validate_final.py: Complete Python code for model training and validation (RF, XGBoost, SVM), SHAP interpretability analysis, and spatial heterogeneity analysis- flood_dataset_normalized.csv: Normalized dataset (972 samples × 10 conditioning factors) used for model training and validation- Table4_model_performance.csv: Model performance metrics (Table 4)- Figure4_shap_importance.csv: SHAP global feature importance (Figure 4)- shap_values_all_samples.csv: Full SHAP value matrix for all 972 samples- susceptibility_predictions.csv: Flood susceptibility predictions for all samples- Links_to_each_dataset.docx: Download links for all original spatial data sources (ASTER GDEM, ESA WorldCover, CHIRPS, OpenStreetMap)