Developing a Forecasting model for uranium occurrence in GII, Northeastern Desert, Egypt using artificial neural networks

Abstract

In the resources sector, artificial neural networks (ANNs) are becoming more and more well-liked. Using datasets of uranium occurrence as input data, ANN technology offers answers to problems. In this paper, a new artificial neural network (ANN) model and Triangulation Irregular Network (TIN) were used for forecasting of uranium occurrence in Gattar II (GII) area, Northeastern Desert, Egypt. The multilayer perceptron ANN model was trained with the Levenberg–Marquardt algorithm, for calculating the uranium (U) occurrence in GII area based on 185 datasets. TIN method showed a clear distribution for uranium ore grade, total gamma ray (total γ-ray) and thorium (Th) content at the studied area. The proposed ANN model achieves coefficient of determination (R2) of 0.993, root mean square error (RMSE) of 67.197%, mean relative error (MRE) of −2.66%, and mean absolute relative error (MARE) of 12.54%.