The use of a multilayer feedforward neural network for mineral prospectivity mapping

The paper reports on a pilot study on the use of an artificial neural network in conjunction with a geographic information system (GIS) for the integration of large multi-source data sets used in regional mineral exploration and prediction of mineral prospectivity. A multilayer feedforward neural network, trained with the error-back-propagation algorithm, is used to estimate the favourability for gold deposits from a raster GIS database for the Tenterfield 1:100000 sheet area, NSW (Australia). To validate and assess the effectiveness of the neural network method, mineral prospectivity maps were also prepared using the empirical weights of evidence and the conceptual fuzzy logic methods. The neural network produces a geologically plausible mineral prospectivity map, similar but superior to the fuzzy logic and weights of evidence maps. The results of this study indicate that neural networks have several advantages over existing methods.