advangeo® Prediction Software 2.2 released

22nd January 2016, Freiberg (Germany)

   

Beak Consultants GmbH announces the release version 2.2 of its advangeo® Prediction Software. New software features include a reviewed data management concept, new model quality evaluation features and the beta version of two additional predictive tools: fuzzy logic and weights of evidence.

Copying of predictive models and their datasets is now as easy as using MS Windows Explorer: Simply copy the entire project folder to any internal or external drive and connect it with your copy of the advangeo® Prediction Software.

Figure: Copying and importing of predictive models with advangeo® Prediction Software

For predictive model evaluation, there are three independent features implemented: the networks weights analysis tool, the model results statistics tool, and the error curve. The network weights analysis tool quantifies the share/contribution of input data layers. It identifies crucial input data and thus helps to understand key model controlling factors. In case of mineral predictive mapping, this tool helps to identify ore controlling features, essential for improving exploration strategies and approaches. The model results statistics tool analyses the statistical distribution of model results and helps to evaluate the model’s reliability. Separate statistics are provided for “true” and “all” values of a model indicating the accuracy of model adaptation. Traditionally, the network error curve is available as an excellent instrument for evaluating the general adaptation of the model. Usually, an error <0.2 indicates significant relationships.

      

Figure: Network weights analysis tool, model results statistics tool and network error curve of advangeo® Prediction Software

In the last years, advangeo® Prediction Software based mineral potential maps have been created for famous mineral provinces, including the gold belts of Tanzania, the greenstone belts of South-West Ghana and the Sn-W province of the German Erzgebirge. Depending on the nature of input data and the size of the area, the resolution of the data sets is usually 50 – 250 m satisfying highest requirements of exploration targeting.

The Gold Potential Map of Tanzania was completed by using advangeo® Prediction Software as a joint research activity between the Geological Survey of Tanzania and Beak Consultants GmbH. It is published in the “Explanatory Note to the Minerogenic Map of Tanzania”. The true resolution of the data is 250 m. The following map extract of the Gold Potential Map shows the famous Londoni area. For more information, please see www.gmis-tanzania.com, www.gst.go.tz and www.beak.de.

   

Figure: Gold Potential Map of Tanzania with detailed map of the famous Londoni area

The South-West Ghana Gold Potential Map was compiled as a joint research project together with the Geological Survey Department of Ghana and the Technical University of Freiberg. The true data resolution of the map is 50 m, which allows identifying details of many gold potential areas. The following map extract of the Gold Potential Map shows the Kibi Belt area.

   

Figure: Gold Potential Map of SW-Ghana with detailed map of Kibi belt area

Within the traditional Sn-W province of the German Ore Mountains (Erzgebirge), hundreds of occurrences are known. Mineral predictive models help to classify known occurrences with regard to their vertical extension and erosion cut. Mineral exploration targets have been identified and recommended for further processing. The resolution of the data is 50 m.

   

Figure: Tin and Tungsten Potential Map of German Ore Mountains (Erzgebirge) with detailed map of westernmost part of the area

New data analysis and predictive tools include fuzzy logic and weights of evidence. Currently, the beta version is available and will be provided free of charge with the software package to any new and existing user. Fuzzy logic is a knowledge driven approach. It provides an interesting alternative to the data driven neural network tool and the weights of evidence tool that was implemented because of its broad use by the explorer’s community. Weights of evidence have been implemented as a data analysis tool and as a data preparation tool for further processing with the neural networks algorithm.

Figure: Fuzzy logic tool of advangeo® Prediction Software

For general information, please see www.beak.de and www.advangeo.com. For software inquiries, exploration targeting support and consultancy services, please contact andreas.barth@beak.de.