Mineral prospectivity mapping (MPM), like other geoscience fields, is subject to a variety of uncertainties. When data about unfavorable sites to find deposits (i.e., drill intersections to barren rocks) is lacking in MPM using machine learning (ML) methods, the synthetic generation of negative datasets is required. As a result, techniques for selecting point locations to represent negative examples must be employed. Several approaches have been proposed in the past; however, one can never be certain that the points chosen are true negatives or, at the very least, optimal for training. As a consequence, methodologies that account for the uncertainty of the generation of negative datasets in MPM are needed. In this paper, we compare two criteria for selecting negative examples and quantify the uncertainty associated with this process by generating 400 potential maps for each of the three ML methods utilized (200 maps for each criterion), which include random forest (RF), support vector machine (SVM), and k-nearest neighbors (KNC). The results showed that applying a geological constraint to the creation of negative datasets reduced prediction uncertainty and improved overall model performance but produced larger areas of very high probability (i.e., >0.9) when compared to using only the spatial distribution of known deposits and occurrences as a constraint. SHAP values were used to find approximations for the importance of features in nonlinear methods, and kernel density estimations were used to examine how they varied depending on the negative dataset used to train the ML models. Prospectivity models for magmatic-hydrothermal gold deposits were generated using data from the shuttle radar terrain mission, gamma-ray, magnetic lineaments, and proximity to dykes. The Juruena Mineral Province, situated in Northern Mato Grosso, Brazil, represented the case study for this work.

A exploração mineral é um campo crucial e estratégico para o desenvolvimento econômico-tecnológico e o bem-estar social. Apesar da crescente demanda por recursos minerais, descobertas globalmente relevantes tem se tornado cada vez mais raras uma vez que os depósitos rasos e/ou mais óbvios se tornaram escassos. Ferramentas que utilizam o aprendizado de máquina (machine learning) têm apresentado grande potencial, auxiliando geocientistas na solução de problemas em todos os campos, incluindo a descoberta de depósitos minerais. A Província Mineral de Alta Floresta (Mato Grosso - MT), vive um novo ciclo exploratório e mostra potencial mineral ainda não plenamente compreendido. Conta com dados aerogeofísicos, geológicos e geoquímicos em densidade e qualidade satisfatórias para o emprego de técnicas de machine learning. Neste projeto foi conduzido o mapeamento de potencial mineral da Província Mineral de Alta Floresta (PMAF) usando métodos de machine learning (ML), a avaliação da explicabilidade e performance dos modelos treinados e a quantificação de incertezas inerentes a este processo. Dados gamaespectrométricos, magnetométricos, modelos digital de elevação e geoquímica de sedimento de corrente foram utilizados na geração de vetores de exploração que representem fonte, transporte e alterações relacionadas aos depósitos auríferos da província. Mapas de gradiente de magnetismo máximo, distância até lineamentos magnéticos, razões Th/K e U/K, elevação e associações Au-Cu foram derivados a partir dos dados brutos visando modelar as assinaturas dos depósitos da região. Mil modelos de ML foram computados para simular possíveis cenários de favorabilidade aurífera utilizando os algoritmos random forest, gradient boosting, support vector machine e k-nearest neighbors. A área de estudo foi dividida em setores norte e sul, sendo as amostras o primeiro utilizadas||para treinamento e do segundo para teste. Com base na media e variância das predições obtidas foi possível definir áreas onde os modelos foram mais ou menos estáveis e fazer interpretações pautadas no modelo metalogenético da PMAF. As regiões com presença de depósitos e/ou ocorrências minerais conhecidas foram, de modo geral, classificadas como favoráveis à presença de depósitos de ouro. Dentre os produtos deste trabalho, foi produzido um manuscrito que será submetido em revistas internacionais.

Geological maps are an extremely valuable source of information for the Earth sciences. They provide insights into mineral exploration, vulnerability to natural hazards, and many other applications. These maps are created using numerical or conceptual models that use geological observations to extrapolate data. Geostatistical techniques have traditionally been used to generate reliable predictions that take into account the spatial patterns inherent in the data. However, as the number of auxiliary variables increases, these methods become more labor-intensive. Additionally, traditional machine learning methods often struggle with spatially correlated data and extracting valuable non-linear information from geoscientific datasets. To address these limitations, a new architecture called the Spatially Constrained Bayesian Network (SCB-Net) has been developed. The SCB-Net aims to effectively exploit the information from auxiliary variables while producing spatially constrained predictions. It is made up of two parts, the first part focuses on learning underlying patterns in the auxiliary variables while the second part integrates ground-truth data and the learned embeddings from the first part. Moreover, to assess model uncertainty, a technique called Monte Carlo dropout is used as a Bayesian approximation. The SCB-Net has been applied to two selected areas in northern Quebec, Canada, and has demonstrated its potential in generating field-data-constrained lithological maps while allowing assessment of prediction uncertainty for decision-making. This study highlights the promising advancements of deep neural networks in geostatistics, particularly in handling complex spatial feature learning tasks, leading to improved spatial information techniques.