Flood susceptibility mapping and assessment using a novel deep learning model combining multilayer perceptron and autoencoder neural networks

Floods are one of the most destructive natural disasters causing financial dam-ages and casualties every year worldwide. Recently, the combination of data-driven techniques with remote sensing (RS) and geographical information sys-tems (GIS) has been widely used by researchers for flood susceptibility map-ping. This study presents a novel hybrid model combining the multilayerperceptron (MLP) and autoencoder models to produce the susceptibility mapsfor two study areas located in Iran and India. For two cases, nine, and twelvefactors were considered as the predictor variables for flood susceptibility map-ping, respectively. The prediction capability of the proposed hybrid model wascompared with that of the traditional MLP model through the area under thereceiver operating characteristic (AUROC) criterion. The AUROC curve for theMLP and autoencoder-MLP models were, respectively, 75 and 90, 74 and 93%in the training phase and 60 and 91, 81 and 97% in the testing phase, for Iranand India cases, respectively. The results suggested that the hybridautoencoder-MLP model outperformed the MLP model and, therefore, can beused as a powerful model in other studies for flood susceptibility mapping.Validerad;2021;Nivå 2;2021-02-22 (johcin)