In the mineral industry, a lot of data is collected at great expense. That dataset is multidisciplinary and can include physical properties, geochemistry, mechanical properties and much more. Traditionally geoscience analysis is confined to a particular discipline - the geochemists work with the geochemistry data, geophysicists work with the geophysics data and so on. This results in a myopic view and does not leverage the entire geoscience dataset, acquired at tremendous cost.
At KORE we demand more from our data and leverage the entire geoscience dataset by using state of the art machine learning technology. KORE's technique allows our clients to, for example, learn about the mechanical properties of their rock by using their geochemistry dataset. This is immensely valuable as most geochemical databases are vast, whereas mechanical properties – key to mill and mine design – are sparse despite their importance.
The example below is a case where our client needed a density model to constrain an inversion, important to model their deposit. They had downhole density in 23 boreholes, which results in a very coarse density model given the sparseness of the measurements. However they had multi-element geochemistry for all 700 boreholes on the project site. KORE built a machine learning model to predict density from 50 element geochemistry and predicted density on all 700 boreholes resulting in a detailed density model.