Table 4 Modeling accuracy in predicting flood hazard mapping [OA: Overall Accuracy]
Training data (70% of flood occurrence points during 2016–18) | Testing data (30% of flood occurrence points during 2016–18) | Validation data (flood occurrence points during 2019–21) |
|---|---|---|
Random forest (RF) | OA: 82.91% Kappa: 0.66 | OA: 82.36% Kappa: 0.65 |
Classification and regression tool (CART) | OA: 81.9% Kappa: 0.64 | OA: 83.37% Kappa: 0.67 |
Support vector machine (SVM) | OA: 82.23% Kappa: 0.64 | OA: 81.15% Kappa: 0.62 |