Novel Back-Propagation Multilayer Perceptron Approach for 2-D Magnetotelluric Modeling

In this study, the novel artificial neural network(ANN) called back-propagation multilayer perceptron (BP-MLP) algorithm with fully connected architecture is proposed to simulate and predict the apparent resistivity and impedance phase of 2-D Magnetotelluric forward modeling. The experiments of this algorithm are made on various benchmark models. The experimental results showed that the proposed neural network is efficient and provides accurate predictions with an average relative error of apparent resistivity and impedance phase less than 1% and 0.5%, respectively. The algorithm's feasibility suggests that the ANN approach provides excellent accuracy results and can be practically used for solving 2-D magnetotelluric modeling.