Geometallurgical Responses on Lithological Domains Modelled by a Hybrid Domaining Framework

Identifying mineralization zones is a critical component of quantifying the distribution target minerals using well-established mineral resource estimation techniques. Domains are used to define these and can be modelled techniques such as manual interpretation, implicit modelling, advanced geostatistical methods. In practise, domaining commonly exercise that labour-intensive prone subjective judgement errors, resulting in largely deterministic output ignores significant uncertainty associated with domain interpretation boundary definitions. Addressing issues requires an objective framework automatically domains quantify uncertainty. This paper presents comparative study PluriGaussian Simulation (PGS) Hybrid Domaining Framework (HDF) based on simulated assay grades XGBoost, machine-learning classification technique trained lithological properties. The two approaches assessed basis boundaries produced data from Iron Oxide Copper Gold deposit. results show proposed HDF accommodate complex multiclass problems imbalanced features. Geometallurgical models Net Smelter Return grinding time demonstrate effectiveness HDF. addition, preprocessing step involving noise filtering method improve performance ML classification, especially cases where difficult predict due similarity geological characteristics inherent data.