Assessment of gold mineralization prospectivity in covered areas using machine learning and artificial intelligence methods (case study of the Sardara depression)
Annotatsiya
Abstract. The article presents a methodology for applying artificial intelligence (AI) and machine learning (ML) technologies to assess the gold mineralization potential of closed areas located between the Southern Tamdytau and Aristantau mountains (Sardara Depression). The obtained results are also presented. Based on digitized and analyzed factual data from geological and geophysical studies conducted at reference deposits, an integrated predictive model was developed to identify prospective sites within adjacent closed territories. During the preliminary processing of geophysical data, various transformation methods were applied, including RTE, ASA, TDR, THDR, UP, and VD, as well as filters such as HP, LP, CET, Frangi, Sobel, and Laplace. For the first time under the conditions of Uzbekistan, a methodology for detecting mineralization zones was developed using machine learning models Random Forest and XGBoost, which made it possible to determine promising gold mineralization zones within closed areas. The obtained results, presented in the form of predictive maps, ensure high reliability in identifying prospective gold-bearing zones and can serve as a basis for further exploration activities. Keywords: Artificial intelligence, machine learning, gold mineralization, geophysical analysis, Sardara depression, prospectivity model, covered areas.