Machine learning for the prediction and Analysis of Bouguer Gravity anomaly of the Kiri uplift in Congo sedimentary Basin

Bouguer gravity anomalies (BGA) play an important role in exploration of mineral resources. Allowing the delineation large geological structures, BGA participate into discovery deposits. However, Kiri uplift region where several oil seeps have been recognized faces sparse coverage data due to difficult conditions acquisition on field. This situation increases non-uniqueness and nonlinearity problems solution using inverse methods. Although, potentially good at quantifying uncertainties, approaches involve enormous computational tasks. We used machine-learning algorithms predict analyze region. The learned perform as a multiple regression. During training steps, each independent variable included X Y coordinates, digital elevation model (DEM) geology. values calculated by experts were provided dependent variables. K-fold cross-validation has ensure models are well fit. Since well-trained should result small losses errors, we experimented optimizers. By comparison, testing results showed that deep neural network-based algorithm (DNN) proven be most efficient with 5.37 Mean Squared Error 1.75 Absolute metrics. DNN accurate prediction, which, together measured reported strongest Pearson correlation coefficient 0.996. In addition, analysis is one conforms perfectly regional geology information study area. Machine learning proved their effectiveness area.ML measurements lackedThe joint predicted, lithology meets expectationsAs predictors X, Y, DEM and, lithology, can customize grids