Modeling of Turkey-Manyas region using Forced Neural Networks

Ali Albora, Onur Osman
In this study, we perform Forced Neural Network (FNN) method to find the parameters of a model according to gravity anomaly profile. Geological modeling is a promoting method to find the shape and location of underground structures. In this method first, one neuron is used and back propagation algorithm is applied to find the density difference. Then, density differences are quantized and mean square error of the model is computed. This process is stopped when the mean square error of the model
more » ... is small enough. FNN method is applied to synthetic data, and then Marmara-Manyas basin map in Turkey. As a result, anomalies of the cross section and the model obtained by FNN are very close to each other.