There is a lot of expensive data in the mining industry and most companies do not extract the full value from it. KORE's cluster analysis process rapidly improves deposit understanding by efficiently identifying, investigating, and exploiting trends in complex multivariate data. Accelerated and increased understanding of complex datasets are achieved through a combination of data validation, machine learning, cluster analysis, traditional statistics and advanced data visualizations. Key outputs of the process include robust, unbiased, quantitative classification schemes, and the development of proxy relationships between traditionally isolated datasets.
KORE's cluster analysis classifies the rock into unbiased domains based on validated quantitative data, including petrophysics and multi-element geochemistry. The domains are visualized (3-D), and iterations, interrogations and comparisons to various classification schemes can be performed rapidly. The ability to generate multiple iterations of project scale petrophysical and geochemical domains and interactively compare them in spatial context to known lithology, alteration, structure and mineralization can lead to valuable insights and better geologic models.
Integrating multi-disciplinary data enables the development of predictive relationships between traditionally isolated datasets. Modern data science techniques applied to in situ data can provide robust real time estimates for traditionally expensive, slow, and sparsely sampled laboratory analyses. KORE has accurately predicted recoverable grade and key assay and geotechnical parameters directly from petrophysical data. This permitted the client to have real-time estimates of recoverable grade from individual boreholes, without requiring additional expensive and time consuming metallurgical tests.