PREDICTIVE TARGETTING

just because it is ai, doesn't mean it is necessarily good....

Unsupervised and supervised classification methods have been around for a while. In fact, earlier version of the ER Mapper software already included classification schemes as early as 1995. Hernan coded an early Neural Net predictive targetting tool in Matlab in 2003. This was later commercialized by Geosoft as a "Predictive Targetting" module. However, the key is the geological treatment of the data.

McGuaig and Hrontsky (2014) established that all mineral systems comprise critical elements in nested scales in space and time:

• Lithosphere architecture
• Favourable geodynamics
• Fertility
• Deposition, Preservation and Modification of a primary depositional zone

The work that we have been doing with Aldo involves compiling all the available data (geophysics, geology, geochemistry, etc), interpret it, and transform that into layers within the Mineral Systems approach. Therefore, rather than using a conductivity map from an EM survey, we assess the presence of magma sources that could give an idea on fertility; and instead of just mapping lineaments out of geophysical data, we identify favourable pathways for mineralizing fluids.