Sztuczne sieci neuronowe w zastosowaniu do budowy modelu geologicznego Pradoliny Głogowsko-Barudzkiej w okolicy Nowej Soli

This article attempts to determine the course of geological strata in part of the G³ogów-Baruth Ice-Marginal Valley, using the multilayer neural networks of a sigmoidal type, trained by the back propagation method that is based on the known theory of gradient optimization methods. Determination of the trend of individual layers was made based on the coordinates (x, y) and the following data: depth of the top, depth of the base and thickness of the strata, obtained from field studies.