Landslide hazard zoning using artificial neural network models and TOPSIS downstream of Sanandaj Dam

The purpose of this study was to investigate and compare the two models of artificial neural network and TOPSIS model in the risk of landslide in the downstream of Sanandaj dam. This was done using ARCGIS software, Python programming language, and artificial neural network models and TOPSIS. For this purpose, 9 input layers were used in landslide risk zoning. Landslide and non-slip points in the area were determined using satellite imagery. Internal weighting was used to determine the weight of the layers. In the neural network model, the data were trained using a multilayer perceptron network with the Adam learning algorithm. The network structure has 9 neurons in the input layer, 30 neurons in the middle layer and 1 neuron in the output layer. In the Topsis model, after declassifying the decision matrix, Shannon's entropy method was used to weigh the criteria and to determine the relative distance from the Euclidean distance. After preparing the models, the study area was analyzed with 970 Km2 with 9 input variables that were converted to raster data into 30x30 pixels. The results of the analysis were mapped with five floors of landslide risk for each model. After applying 5 methods of calculating the error rate to validate the models, it was found that the perceptron neural network model has less error and more adaptation and is better compatible with the geography of the region.