Prediction of Gold associated Mineral worth: An application of mathematically driven artificial neural network technique

Abstract

The elemental composition of other associate minerals existing with gold is a significant asset that defines the amount of additional economic contribution that can be obtained from the gold tailings. The elemental composition is a needed factor in increasing the economic value of gold run-off and getting a clear estimation for the quantity of value-added elements in each tonne of gold sand scooped during the separation process. In this study, the artificial neural network (ANN) modeling technique was used to develop an economic worth prediction model for 10 gold-associated minerals. The developed models have a 1:7:10 architecture and were trained using the ANN Bayesian regularization training algorithm. According to the root mean square error values, the results revealed that the predicted values of the associated minerals are closer to the measured values. Also, the developed model prediction performance was found to be appropriate for the estimation of gold-associated mineral economic benefits based on the high coefficient of determination and variance account. The model performance evaluation results show that the developed ANN models are suitable for economic estimation of gold-associated mineral worth.