A 50-element stream sediment geochemistry survey with a sample density of one sample per km² shows anomalies for a number of high technology metals, including Sn, W and the battery metals Co and Li. Apart from the anomalies associated with known deposits, low-level anomalies of Li and Co are widespread in anchimetamorphic to greenschist facies metasedimentary rocks. Li anomalies are prominent in Phycoden Group phyllites in the Vogtland and parts of the Frauenbach and Thum Groups in the West Erzgebirge Transverse zone. Co is enriched in the Brunndöbra Subformation (Klingenthal Group) where it is probably associated with Besshi-type massive sulfide mineralisations.
We use self-organizing maps (SOM), a type of artificial neural network (ANN), for data analysis and data fusion with geophysical and structural-lithologic data to identify areas of interest for further exploration. Using known deposits as training data, the SOM-transformed data are converted to mineral predictive maps with a multi layer perceptron, a different type of ANN. This combined machine learning approach overcomes some of the problems in applying ANNs to mineral predictive mapping, in particular the problem of imbalanced training data.
The paper has been compiled in the frame of "NEXT - New EXploration Technologies" project. This project has received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 776804.
Andreas Brosig, Andreas Barth, Peggy Hielscher, Claus Legler, Stefan Schaefer, Peter Bock, Andreas Knobloch
Beak Consultants GmbH, Germany