Integration of mineral potential maps from various geospatial models

Abstract

The purpose of this study is to map mineral potential in Gangreung area, Korea. The mineral potential map was made and validated using likelihood ratio, logistic regression and artificial neural network models with a geographic information system (GIS). Moreover integration of the models has applied to get the better accuracy than each model. For this, the factors related to Au-Ag mineral occurrence were compiled in the GIS database. The factors are the geological data of lithology and fault structure, geochemical data. Using these factors, the potential of mineral were analysed using the 3 models. The validation result showed that the likelihood ratio, logistic regression and artificial neural network models had 83.70%, 81.91% and 77.37% accuracies. But the integrated mineral potential map, prediction accuracy was 92.94%. The generated maps could be used to not only predict known areas of Au-Ag occurrence, but also identify areas of potential mineralization where no known deposit occurs.