Mineral Resource Estimation Research Articles

Advancing Iron Ore Grade Estimation: A Comparative Study of Machine Learning and Ordinary Kriging

Mineral grade estimation is a vital phase in mine planning and design, as well as in the mining project’s economic assessment. In mining, commonly accepted methods of ore grade estimation include geometrical approaches and geostatistical techniques such as kriging, which effectively capture the spatial grade variation within a deposit. The application of machine-learning (ML) techniques has been explored in the estimation of mineral resources, where complex correlations need to be captured. In this paper, the authors developed four machine-learning regression models, i.e., support vector regression (SVR), random forest regression (RFR), k-nearest neighbour (KNN) regression, and extreme gradient boost (XGBoost) regression, using a geological database to predict the grade in an Indian iron ore deposit. When compared with ordinary kriging (R2 = 0.74; RMSE = 2.09), the RFR (R2 = 0.74; RMSE = 2.06), XGBoost (R2 = 0.73; RMSE = 2.12), and KNN (R2 = 0.73; RMSE = 2.11) regression models produced similar results. The block model predictions generated using the RFR, XGBoost, and KNN models show comparable accuracy and spatial trends to those of ordinary kriging, whereas SVR was identified as less effective. When integrated with geological methods, these models demonstrate significant potential for enhancing and optimizing mine planning and design processes in similar iron ore deposits.

Some basic improvements in mineral resource estimation and reporting

ABSTRACT Bulk density miscalculations and subsequent reporting can produce resource estimation errors and unexpected financial outcomes. Density and bulk density vary throughout a mineral deposit, especially in porous deposits and deposits containing variable amounts of high-density minerals (e.g., barite). Resource estimators must correlate bulk density with core recovery and apply quality assurance/quality control analysis to bulk density data. Core recovery is an important factor in resource estimation and reporting. Block tonnages (ore and waste) are estimated from volumes using a dry bulk density value. Ore dilution is a volumetric effect integral to ore reserve estimation that also has the potential to affect the economics of mining a deposit. Both the bulk density and dilution problems could be overcome by reporting a mineral resource estimate in grade per unit volume.

Byproduct-to-Host Ratios for Assessing the Accessibility of Mineral Resources.

Mineral resources are essential for reaching net-zero ambitions by 2050. There is a rising diversity of metals in electricity generation and storage technologies, as well as for mobility technologies. However, little is known about the future supply of minor elements historically mined in low volumes such as indium, tellurium, germanium, or tantalum. Those minor elements are found in lower concentrations in the ores of major elements and therefore rarely form economic deposits on their own. Such elements are often produced as byproducts of a host (or "target commodity", which underpins the bulk of a mine's profitability) in ore, e.g., in porphyry ore, tellurium is a byproduct where copper is the host. As a result, the primary supply of those minor elements depends on the supply of the major elements. Such dependency has not been accounted for in scenarios of the mineral supply. To address this gap, we developed a methodology to harmonize scattered data of mineral resource estimates and to calculate the mass ratio between the byproduct and the host in ores and concentrates, called the byproduct-to-host (BtH) ratio. We collected crude ore tonnage and element grades, among other key data, from the state-of-the-art literature and publicly available mining company reports. Our data set covers 3422 deposits across 141 countries providing 22 275 BtH ratios. The future supply of minor elements can be derived by multiplying the primary production of host elements by the developed BtH ratios, noting the limitations of data representativity. The open-access nature of this work facilitates the enrichment and update of this data set in the coming years.

Impact of Competent Persons' judgements in Mineral Resources classification

Uncertainty with regard to estimated grades and tonnages of a mineral deposit demands risk assessment in order to build investor confidence and attract the interest of other stakeholders in the success of a project. Uncertainties associated with Mineral Resource estimates can lead to unreliable production schedules and unpredictable cash flows. However, the techniques used in the mining industry to determine these uncertainties are inconsistent, because the important decisions taken in the process are solely dependent on the responsible Competent Person (CP), without limitations. This leads to disparities between different CPs' results, using data-sets from the same drill-holes. The various standard codes for public disclosure provide guidelines and recommendations for the classification of Mineral Resources and Reserves but do not provide details on, for example, the amount of geological and geostatistical information needed to qualify for each category of Resources and Reserves. The parameters used to generate the classification categories are subjectively assumed by the responsible CP. In this paper we investigate the impacts of different CPs' judgements on resource classification, using the same data-sets. The results underpin the need for the mining industry to develop a uniform Mineral Resources and Reserves classification framework that can minimize or avoid significant discrepancies that lead to potential misleading public disclosures.

Compositing and regularization of drillhole data for geostatistical resource estimation

Compositing or regularization of drillhole data is common practice in mineral resource estimation and is deemed a necessary step in producing unbiased estimates of mineral resources and reserves. Commonly, data are collected over irregular distances due to the varying relative thicknesses of lithologies drilled or sampling/assaying strategies. This necessitates data transformation to regular lengths of equal size to ensure that all data have the same sample support. However, there have been few detailed publications on the effect of this process on the composited data that are subsequently taken forward for the estimation process. In this paper, three currently available compositing methods are reviewed and the effects of inappropriate compositing methodologies presented. It is shown through a case study that compositing samples to different lengths leads to changes in the average and variance of the grades in the drillcores in the dataset, which will impact the final estimated value. These differences are exacerbated by breaks or gaps in data where, for a variety of reasons, there has been no data collection or data have been lost. The importance of appropriately treating blank and zero data is also presented. Globally, these differences might be minimal, but locally may be substantial, affecting the efficiency of the estimation and subsequent use of the results in, for example, mine planning and reconciliation. Further detailed investigation of compositing practices is required if the full implications of compositing are to be understood and any induced bias effectively defined.

DESARROLLO DE MÉTODOS BASADOS EN REDES NEURONALES EN LA ESTIMACIÓN DE RECURSOS MINERALES

Due to economic and physical limitations, our understanding of mineral resources in a specific area of interest is limited and fragmented. Traditionally, this problem has been solved using the Kriging geostatistical method, where the ore grade is estimated at unmeasured locations using known values of the grade at surrounding points. The advantage of this method lies in the calculation of weights through a spatial variability model known as a variogram. However, the method is imperfect, as it is based on the assumption of stationarity, aditivity, linearity and potential subjectivity in variographic modelling. This study proposes to approach the mineral resource estimation problem as a regression problem using neural networks, which are not subject to the restrictions of stationarity, aditivity, linearity and spatial modelling of geostatistics methods. Kriging and a radial basis function neural network and a multilayer perceptron have been compared using different validation metrics. The results show that a properly trained neural network model, with appropriate labelling of the mineral grade and its input characteristics, achieves similar results to the geostatistical approach, with a significant reduction in time, while avoiding all the aforementioned assumptions. However, neural networks do not consider the spatial correlation of ore grade or reproduce it at the locations where it was measured, characteristics that have marked distrust in its industrial implementation and that are discussed in this article, finally proposing an adjustment between both approaches at a minimum sacrifice of time and labor costs. Keywords Neural networks; machine learning; geostatistics.

Practical applications of quality assurance and quality control in mineral exploration, resource estimation and mining programmes: a review of recommended international practices

Independent quality assurance and quality control (QA/QC) programmes are required by reporting codes for publicly listed companies and are necessary to optimize data quality at all stages of the sampling, preparation and analytical processes involved in mineral exploration, resource estimation and mining grade control. QA/QC programmes should be adjusted over time to meet changing requirements in data quality at different stages of mineral resource development and exploitation. Certified reference materials are used to monitor accuracy and bias at the project laboratory relative to consensus values for the material from round-robin certification analyses. They are also used to monitor drift over time within an individual laboratory and to identify significant failures in QC at the analytical batch level caused by abrupt changes in concentration related to re-calibration of instruments or procedural changes at the laboratory. Duplicate analyses of sample material are generated at key stages of sampling and preparation to estimate the precision of data generated at each stage. Invariably, the largest source of uncertainty occurs during the initial sampling. Coarse blanks are used to monitor cross-contamination between samples or from sample preparation equipment. Furthermore, each of these QC sample types can be used to discover possible sample mix-ups. Supplementary material: An Excel spreadsheet for the calculation of the average coefficient of variation (CV AVE ) (Appendix C) is available at https://doi.org/10.6084/m9.figshare.c.7070137 Thematic collection: This article is part of the Reviews in Exploration Geochemistry collection available at: https://www.lyellcollection.org/topic/collections/reviews-in-exploration-geochemistry

Some common flaws encountered in mineral resource estimation and how to avoid them

ABSTRACT Preparation of a mineral resource estimate (MRE) is an essential component in the mining cycle, as errors that occur in an MRE will affect all following steps that rely upon its accuracy. Over the course of many decades, SLR Consulting (Canada) Ltd. and predecessor Roscoe Postle Associates have observed a number of common errors that occur at all stages of the workflow. The purpose of this paper is to share some of SLR’s experiences relating to the errors encountered during the preparation of MREs and to present some solutions for avoiding these errors. SLR observes that the source of many of the flaws is the result of the level of knowledge, experience, judgment, or expertise by the practitioner of the fundamental principles of mineral resource estimation and with the software package used in preparing the MRE. Attention to detail and adherence to high quality standards throughout the estimation process is the first step in avoiding many of the errors. A critical item for all practitioners to bear in mind is that they are accountable and bear the ultimate responsibility for all aspects of their work.

Estimation of gold-bearing mineral resources and reserves in North-Eastern Yakutia

Estimation of gold-bearing mineral resources and reserves in North-Eastern Yakutia

Augmenting Stationary Covariance Functions with a Smoothness Hyperparameter and Improving Gaussian Process Regression Using a Structural Similarity Index

Gaussian process (GP) regression provides a probabilistic framework for modeling geochemistry in mineral resource estimation and environmental monitoring applications. An issue with this approach is that the kernel hyperparameters obtained by maximizing the log-marginal likelihood (LML) often produce GP posterior mean estimates that are overly smooth. This motivates the development of augmented kernels that are more capable of capturing the variability inherent in geological/geochemical processes than the existing stationary covariance functions like the Matérn kernels. This paper makes two contributions. First, it describes an extended class of stationary kernels that contain an extra smoothness hyperparameter (α\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\alpha $$\\end{document}) which can be learned from the input data. Valid intervals for α\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\alpha $$\\end{document} that lead to positive semi-definiteness are determined. Second, it uses a statistical measure called the structural similarity index (SSIM) to quantify smoothness and the spatial fidelity of GP solutions with respect to the input samples. This provides a new way for validating and optimizing GP models. Statistical and spectral analyses provide insights into the behavior of α\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\alpha $$\\end{document} in the augmented kernels which retain useful properties such as sparsity. Results on the Northern Great Basin geochemical dataset demonstrate that, all things being equal, (1) adjusting α\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\alpha $$\\end{document} increases spatial fidelity in GP regression; (2) SSIM is a more reliable spatial quality measure than LML; and (3) the optimal α\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\alpha $$\\end{document} value obtained is correlated with the Wiener entropy of the random process, which indicates the spectral flatness of the chemical signal in the Fourier domain. For GP regression, function smoothness may be defined using Sobolev space. The results show that α\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\alpha $$\\end{document} regulates over-smoothing by moderating the rate of decay in the power spectrum of the equivalent kernel.Graphic abstract

Information Fusion and Metallogenic Prognosis of Gold Deposits in the Qixia Area, Northern Shandong Province, China

Analyzing and fusing information layers of exploratory parameters is a critical step for enhancing the accuracy of identifying mineral potential zones during the reconnaissance stage of mineral exploration. The Qixia area in Shandong Province is characterized by intricate geological structures and abundant mineral resources. Numerous gold polymetallic deposits have been discovered in this region, highlighting the potential for discovering more such deposits in the ore concentration zone and its adjacent areas. In this study, we focus on the Qixia area and employ the box dimension method to analyze the fractal dimension of fault structures. We investigate the relationship between orebody occurrence and fault incidence within the mining region. Furthermore, we combine fractal analysis with Fry analysis to comprehensively predict the metallogenic potential in the area. This study reveals the fractal dimension values of fault structures, demonstrating that fault structures govern the distribution of ore bodies, with NE and NW fault structures being the primary ore-hosting features. Based on thorough analysis, we hypothesize that gold deposits in this area are generally distributed along the northeastern direction. By considering mineral distribution characteristics, this study identifies five potential metallogenic prospect areas within the study region. Capitalizing on advancements in information technology and big data, digital geology has gained prominence in prospecting and prediction. To this end, we construct a multi-information comprehensive prospecting model based on the structure-geochemical anomaly-mineralization alteration, employing the convolutional neural network (CNN) model for quantitative estimation of regional gold mineral resources. The findings validate the CNN model’s robust prediction performance in this area, leading to the determination of five prediction prospects. We observe a notable congruence between the two methods, offering significant insights for subsequent exploration endeavors in the region.

Addressing Geological Challenges in Mineral Resource Estimation: A Comparative Study of Deep Learning and Traditional Techniques

Spatial prediction of orebody characteristics can often be challenging given the commonly complex geological structure of mineral deposits. For example, a high nugget effect can strongly impact variogram modelling. Geological complexity can be caused by the presence of structural geological discontinuities combined with numerous lithotypes, which may lead to underperformance of grade estimation with traditional kriging. Deep learning algorithms can be a practical alternative in addressing these issues since, in the neural network, calculation of experimental variograms is not necessary and nonlinearity can be captured globally by learning the underlying interrelationships present in the dataset. Five different methods are used to estimate an unsampled 2D dataset. The methods include the machine learning techniques Support Vector Regression (SVR) and Multi-Layer Perceptron (MLP) neural network; the conventional geostatistical methods Simple Kriging (SK) and Nearest Neighbourhood (NN); and a deep learning technique, Convolutional Neural Network (CNN). A comparison of geologic features such as discontinuities, faults, and domain boundaries present in the results from the different methods shows that the CNN technique leads in terms of capturing the inherent geological characteristics of given data and possesses high potential to outperform other techniques for various datasets. The CNN model learns from training images and captures important features of each training image based on thousands of calculations and analyses and has good ability to define the borders of domains and to construct its discontinuities.

Towards quantifying uncertainties in geological models for mineral resource estimation through outside-in deposit-scale structural geological analysis

Mineral Resource downgrades can be prevented by integrating better structural geological interpretations into project evaluations. Geological modelling for Mineral Resource estimation is presented as a systematic process of structural geological analysis at the deposit-scale using drill-sampled grade data. The initial step in this process is to use Maximum Intensity Projection (MIP) rendering to interpret axial symmetry and its orientation from the grade interval mid-points at the deposit-scale, thus reducing the modelling process’s degrees of freedom to only three—estimating the prolateness of the grade or lithological distribution. This approach contrasts with the brute-force methods for assessing modelling uncertainty because the axial direction is fixed to the down-plunge orientation, thus eliminating from the outset, structurally implausible geometric configurations and unrealistic geological scenarios common in brute-force methods. Adopting this new approach of creating geological models means moving away from conventional mining industry methods that prioritise operational considerations in constructing resource estimates but fail to explain the geometric complexities of mineral deposits. We use experiments and a practical case study of a historical gold mine in Western Australia to highlight the significance of analysing the structural control of a gold deposit at the deposit-scale using raw grade data. Our research is based on anecdotal evidence suggesting that erroneous geological modelling leads to unforeseen Mineral Resource downgrades. However, we cannot confirm this hypothesis because there have been few, if any, independent public inquiries into Mineral Resource downgrades. To establish the underlying reasons for significant resource downgrades with certainty, impartial investigations (modelled on accident investigations in the aviation industry) are imperative.

Application of Artificial Neural Network for the Prediction of Copper Ore Grade

Precise prediction of ore grade is essential in feasibility studies, mine planning, open-pit and underground optimization, and ore grade control. Conventional methods, such as geometric and geostatistical methods, are the most popular techniques for mineral resource estimation but fail to capture the complexity of orebodies. Due to this limitation, grades are incorrectly estimated, leading to inaccurate mine plans and costly financial decisions. Here, we propose an ore grade prediction method using an artificial neural network (ANN). We collected 14,294 datasets from the Jaguar mine in Western Australia. The proposed model was developed by incorporating lithology, alteration, eastings, northwards, altitude, dip, and azimuth to predict the grade, and the performance evaluation metrics were measured based on the mean absolute error (MAE), mean square error (MSE), root mean square error (RMSE), correlation coefficient, R, and coefficient of determination (R2). The proposed ANN model outperformed classic machine learning methods with R2, R, MAE, MSE, and RMSE of 0.584, 0.765, 0.0018, 0.0016, and 0.041, respectively. The Shapley technique was used to evaluate the feature importance of the input variables for the grade prediction. Lithology demonstrated the highest influence on ore prediction, whereas eastings had the least impact on output. The proposed approach is promising for ore model prediction.

Application of the k-Prototype Clustering Approach for the Definition of Geostatistical Estimation Domains

The definition of geostatistical domains is a stage in the estimation of mineral resources, in which a sample resulting from a mining exploration process is divided into zones that show homogeneity or minimal variation in the main element of interest or mineral grade, having geological and spatial meaning. Its importance lies in the fact that the quality of the estimation techniques, and therefore, the correct quantification of the mineral resource, will improve in geostatistically stationary areas. The present study seeks to define geostatistical domains of estimation for a mineral grade, using a non-traditional approach based on the k-prototype clustering algorithm. This algorithm is based on the k-means paradigm of unsupervised machine learning, but it is exempt from the one-time restriction on numeric data. The latter is especially convenient, as it allows the incorporation of categorical variables such as geological attributes in the grouping. The case study corresponds to a hydrothermal gold deposit of high sulfidation, located in the southern zone of Peru, where estimation domains are defined from a historical record of data recovered from 131 diamond drill holes and 37 trenches. The characteristics directly involved were the gold grade (Au), silver grade (Ag), type of hydrothermal alteration, and type of mineralization. The results obtained showed that clustering with k-prototypes is an efficient approach and can be used as an alternative or complement to the traditional methodology.

Review of machine learning-based Mineral Resource estimation

Mineral Resources estimation plays a crucial role in the profitability of the future of mining operations. The conventional geostatistical methods used for grade estimation require expertise, understanding and knowledge of the spatial statistics, resource modelling, geology, mining engineering as well as clean validated data to build accurate block models. However, the geostatistical models are sensitive to changes in data and would have to be rebuilt on newly acquired data with different characteristics, which has proved to be a time-consuming process. Machine learning methods have in recent years been proposed as an alternative to the geostatistical methods to alleviate the problems these might suffer from in Mineral Resource estimation. In this paper, a systematic literature review of machine learning methods used in Mineral Resource estimation is presented. This has been conducted on such studies published during the period 1990 to 2019. The types, performances, and capabilities, of several machine learning methods have been evaluated and compared against each other, and against the conventional geostatistical methods. The results, based on 31 research studies, show that the machine learning-based methods have outperformed the conventional grade estimation modelling methods. The review also shows there is active research on applying machine learning to grade estimation from exploration through to exploitation. Further improvements can be expected if advanced machine learning techniques are to be used.

“Reasonable prospects” in mineral resource estimation and reporting

ABSTRACT A fundamental component of the Definition Standards for Mineral Resources is the requirement that the material have “reasonable prospects for eventual economic extraction” (RPEEE). The first adoption of the RPEEE requirement by a widespread group was in 1997 as part of the Denver Accord meeting and has since become an international standard. Mineral resource statements are prepared for mineral properties that are at various stages in the mining cycle, from the discovery stage through to the production stage. The confidence level of information available for the development of the input parameters for the technical and economic components of RPEEE varies greatly among stages. Consequently, the RPEEE assumptions may evolve with the stage of the mineral property as well. The economic aspects of RPEEE are commonly achieved by selecting appropriate input parameters for establishing a cut-off grade. The technical aspects can include consideration of such items as minimum widths, spatial continuity, and the application of appropriate constraining surfaces and volumes. In all cases, the responsibility of ensuring that the mineral resource statements are prepared in accordance with the requirements of the Canadian Institute of Mining, Metallurgy and Petroleum Definition Standards ultimately resides with the qualified person.

Výsledky druhej fázy geologického prieskumu hlbokomorských polymetalických konkrécií v prieskumnom území Spoločnej organizácie Interoceanmetal

The exploration rights of the Interoceanmetal Joint Organization for exploration of polymetallic nodules (PMN) are granted from 29 March 2001 to an area located within the Clarion-Clipperton Zone (CCZ) in the eastern central Pacifi c Ocean. Exploration area covers 75,000 km2 and consists of two sectors (B1 and B2). The B2 sector comprises four exploration blocks (H11, H22, H33 and H44). The most prospective area, selected for detailed research, is marked as H22_NE exploitable block and delineated within the H22 exploration block. The article presents results of geological survey, based mostly on the data collected during the second phase of exploration in the licence area (2016–2021, extension of the contract). Results are based on IOM’s expeditions and relevant analytical work. During the IOM-2018 expedition high resolution bathymetric survey of H11, H22, H33 and H44 exploration blocks was carried out. The IOM-2019 expedition provided a new set of the data obtained using the distance methods (side-scan sonar, profi ler) and contact methods (box-corer and gravity corer) in H22_NE exploitable block, H33 exploration block and preliminary delineated Preservation Reference Zone. The study was focused on analytical work based on sediment and nodule analyses of samples in H22 exploration block and H22_NE exploitation block. New estimation of mineral resources in B2 sector was caried out using the geostatistical method of ordinary block kriging with Yamamoto correction. The polymetallic nodule resources have been classifi ed within the Inferred, Indicated and Measured Resources categories of the CRIRSCO classification system.

Resource Estimation in Multi-Unit Mineral Deposits Using a Multivariate Matérn Correlation Model: An Application in an Iron Ore Deposit of Nkout, Cameroon

Modeling the spatial dependence structure of metal grades in the presence of soft boundaries between geological domains is challenging in any mineral resource estimation strategy. The aim of this work was to propose a structural model adapted to this type of geological boundary, based on a multivariate Matérn model that fits the observed direct (within domain) and cross (between domains) correlation structures of metal grades. The methodology was applied to a case study of an iron deposit located in southern Cameroon. Cross-validation scores show that accounting for the grade correlation across domain boundaries improved the traditional workflow, where the grade was estimated in each domain separately. The scores were significantly better when we also ensured that the mean grade was locally invariant from one domain to another to reflect the grade continuity across the domain boundary.

A hybrid framework for modelling domains using quantitative covariates

Domains define the boundaries of mineralisation zones, within which the grade distribution of the target minerals can be quantified via an established mineral resource estimation procedure. Available domain modelling techniques include manual interpretation, implicit modelling and advanced geostatistical approaches. In mining applications, the most commonly used method is manual domaining, which is labour-intensive and prone to subjective judgement errors. In addition, the output is largely deterministic and ignores the significant uncertainty associated with the domain interpretation and boundary definitions. There is, therefore, a need for a more objective framework that can automatically define mineral domains and quantify the associated uncertainty. This paper describes such a framework, which consists of a hybrid approach based on simulated grade distributions and a machine learning (ML) classification technique, XGBoost, trained on lithological properties. Data from an Iron Oxide Copper Gold (IOCG) deposit are used as a case study to demonstrate the application of the proposed method. The study shows that the approach can handle complex multi-class problems with imbalanced features, and it can quantify the uncertainty of domain boundaries. A noise filtering method is used as a pre-processing step to improve the performance of the ML classification, especially in the case of problematic classes where domain boundaries are difficult to predict due to the similarity in geological characteristics.