Application of support vector machine in modeling land subsidence in parts of Aliabad plain of Qom

Due to the enhancement of urbanization, industry, and agriculture, increasing temperature, and decreasing rainfall, the demand for water supply has increased. Excessive extraction of groundwater for consumption imposes a decline in the groundwater level and the occurrence of subsidence as a consequence. In this study, a support vector machine approach is applied to model the subsidence. Groundwater level drop, the thickness of alluvial sediments, the transmissivity of alluvial sediments, and elasticity modulus have been used as independent parameters of subsidence modeling by a support vector machine. The results indicate that the support vector machine model can model the subsidence with reasonable accuracy. To verify the performance of the support vector machine, the model's results have been evaluated with the measured values of the InSAR method obtained from the satellite images of some parts of Aliabad Plain. In addition, to examine the impact of model input parameters on subsidence, a sensitivity analysis has been conducted, and the results illustrate that the occurrence of subsidence is distinctly dependent on the drop in the underground water level in the region. The generalizability of the model has been investigated by using a new dataset, and the results indicate the generalizability of the subsidence support vector machine model.