3D mineral exploration Cu-Zn targeting with multi-source geoscience datasets in the Weilasituo-bairendaba district, Inner Mongolia, China

The Weilasituo-bairendaba district is located at the eastern end of Central Asian Orogenic Belt, which an important component Cu-Pb-Zn polymetallic metallogenic belt on Western slope Greater Xing’an Range in Inner Mongolia, China. known Cu-Zn deposits such as Weilasituo deposit and Bairendaba Ag-Pb-Zn are same tectonic-magmatic product. district’s structure framework consists NE-trending regional faults, while secondary faults provide channels space for mineralization. ore-bearing rocks either Baoyintu Group gneisses or quartz diorites. typical exhibit obvious Cu, Pb, Zn geochemical anomaly well magnetic anomaly. district-scale two-dimensional (2D) mineral prospectivity modeling has been reported. Nowadays, three-dimensional (3D) necessary urgent. Integrated geology accumulated exploration data, above four criteria (regional fault, susceptibility) used 3D modeling. Filtering (upward continuation, low pass filtering, empirical mode decomposition), inversion techniques were to construct geological models. Excellent machine learning algorithms random forest (RF) XGBoost applied. two methods confirm each other improve accuracy In this paper, repeated sampling Bayesian optimization combined tune This joint method can avoid contingency caused by negative samples, also realize automatic hyperparameters. optimal models (RF28 XGBoost11) selected among thirty training obtained areas under ROC curves RF28 XGBoost11 0.987 0.986, respectively. prediction-area (P-A) plot C-A fractal delineate targets grade targets. divided into Ⅰ-level Ⅱ-level I- II-targets not only highly consistent with deposits, but ore-forming features. beneficial district.