Evaluation, Prediction and Regional Analysis of Floods Using Data Mining Models (Frizi Watershed)

Floods are one of the natural hazards that cause financial and human losses every year. This research was carried out to determine the flood risk zones in the Friesian watershed with Artificial Neural Network Model and Support Vector Machine to identify the influential factors in the occurrence of floods in the region, necessary measures for control, planning and protective and management measures in to reduce the risk of flooding.

This was an applied, analytical and developmental study carried out in 2023 in Firizi watershed. 14 environmental parameters were used for flood risk zoning. Then, 150 flood points in the basin were extracted using satellite images and randomly divided into 70% for model training and 30% for validation. Then, the layers were weighted, and the flood potential maps were classified into 5 categories: Very high, high, medium, low, and very low, using the natural fracture method. Then, the accuracy of the models used in the research was investigated using the Receiver Operating Characteristic method.

The research assessed the accuracy of the models by employing the Receiver Operating Characteristic approach. The collected findings indicate that the Artificial Neural Network model outperformed the Support Vector Machine model, with an Area Under the Curve value of 0.923 compared to 0.898.

The flooding in the basin primarily affected areas with gentle slopes, rocks that are resistant to water flow and not easily penetrated, lower altitudes, and lands located next to rivers (known as floodplains). The basin contains areas that are classified as high and extremely high-risk, primarily in its central and outflow regions.