Geologists at the Karpinsky Institute use artificial intelligence to improve the reliability of forecasts for identifying mineral prospects

Currently, with the development of modern technologies, spatial statistical methods for forecasting mineral deposits have become widespread in the world. They are based on two fundamental approaches: empirical-statistical and genetic.

Methodologies for empirical-statistical predictive metallogenic analysis use GIS software and digital modelling of spatial data based on artificial intelligence technologies (neural networks, machine learning, etc.). The analysis is based on spatial relationships between known mineral deposits or occurrences and specific geological, geochemical, geophysical features represented in digital spatial datasets.

Modelling methods based on a genetic approach differ from data-driven analysis by using conceptual models of ore formation (genetic geological models) rather than empirical relationships. Predictive modelling is implemented in a GIS environment using analysis methods such as index overlay and fuzzy logic.

Predictive metallogenic schemes based on the described methods and approaches make it possible to comprehensively and with maximum efficiency use various sets of geodata, identify new patterns of mineralization localization and identify prospects at a qualitatively new level.

According to Vasily Leontyev, Deputy Director of the Centre for Predictive Metallogenic Research, “Today, the Karpinsky Institute has key competencies both in the field of metallogeny and in the field of geographic information systems. The Institute’s experts are developing new technological solutions based on domestic software, as well as practical recommendations for identifying mineral prospects promising for hydrothermal deposits of non-ferrous and precious metals and diamond deposits based on the use of artificial intelligence technologies.”

The implementation of the developed technological solutions will allow the Institute's geologists to increase the reliability of the forecast and the degree of validity of identifying mineral prospects for medium-scale and large-scale geological surveying and prospecting work for various types of minerals.