Mapping Mineral Prospectivity via Semi-supervised Random Forest

Prospectivity mapping based on deep learning typically requires substantial amounts of geological feature information from known mineral deposits. Due to the limited spatial distribution of ore deposits, the training of predictive models is often hampered by insufficient positive samples. Meanwhile, data-driven mineral prospectivity mapping often overlooks domain knowledge and expert experience, leading to poor interpretability of predictive results. To address this problem, we employed the Gaussian mixture model (GMM) for spatial feature classification to expand the number of positive samples. The approach integrated the embedding of geological map knowledge graphs with geological exploration data to enhance the knowledge constraints of the prospecting model, which enabled the integration of knowledge with data. Considering the complex spatial structure of geological elements, a bi-branch utilizing the 1-dimensional convolutional neural network (CNN1D) and graph convolutional network (GCN) was used to extract geological spatial features for model training and prediction. To validate the effectiveness of the method, a gold mineralization prediction study was conducted in the Wulonggou area (Qinghai province, western China). The results indicate that, when the number of GMM spatial feature classifications was 17, the positive-to-negative sample ratio was optimal, and the embedding of the knowledge graph controlled the prediction area distribution effectively, which demonstrated strong consistency between the prospecting area and the known mineral deposits. Compared with the predictions by CNN1D, the fused prediction model of CNN1D and GCN yielded higher accuracy. Our model identified 11 classes of mineralization potential areas and provides geological interpretations for different prediction categories.

Fuzzy analytical hierarchy (AHP) approach is a method for mineral prospectivity mapping (MPM) that is generally used for mineral exploration. This method is feasible for multi-criteria decision-making (MCDM) issues. Geographical information system (GIS) and fuzzy AHP have been applied in this paper to obtain prospectivity model for Calcic Iron Skarn (CIS) mineralization. Several thematic (such as geological, geophysical, and geochemical) geo-datasets have been collected, analysed and integrated based on fuzzy AHP method, in Varan area, central of Iran. Three professional economic geologists with the experience on exploration of CIS mineralization have been used to allocate appropriate weights to layers. Then fuzzy operator was used to integrate weighted evidence layers and mineral prospectivity map of CIS mineralization in Varan area was generated. Eventually for confirming the accuracy of the applied manner, locations of recognized mineral deposits in the Varan area were compared with the generated mineral prospectivity map. The results presented acceptable responses. For detailed assessment of the CIS mineralization model and due to the fact that geological features have fractal dimensions, we used C-A model for classification of prospectivity model to determine thresholds for the final prospectivity map. C-A fractal model, determine thresholds for classifying values in evidential maps. The final prospectivity map was confirmed by checking field of three target areas.

GIS-based mineral prospectivity mapping (MPM) is a computer-aided methodology for delineating and better constraining target areas deemed prospective for mineral deposits of a particular type. The underlying algorithms are well-established and well-understood, but on the whole, MPM that is a multi-faceted and multi-criteria approach, is faced with a high degree of uncertainty. We distinguish three principal types of uncertainties: (1) data-related (e.g., the sometimes erroneous, inadequate, incomplete, unevenly distributed or poorly resolved nature of the input data); (2) model-related (e.g., the diversity and inherent natural variability of mineral deposits, our lack of complete knowledge of the targeted mineral deposit type, and our imperfect ability to interpret geoscience datasets); and (3) judgment-related (e.g., the influence of cognitive heuristics and biases). In this contribution, we review and characterize the key uncertainties listed above and provide possible solutions as to how they may be recognized and mitigated in the context of MPM. This review also clearly illustrates the need for future studies designed to carefully monitor each step of the MPM process and aims at reducing uncertainty by, for example, (1) using carefully vetted, high-quality input data, (2) developing targeting models based on the best possible understanding of the underlying mineral deposit models and backed by machine learning-based simulations of likely ore-forming processes, and (3) adopting advanced methods such as deep learning algorithms for effective integrating of predictor maps.

Mineral Prospectivity Mapping (MPM) is a crucial process in mineral exploration, traditionally hampered by subjective interpretations and labor-intensive methods leading to unreliable outcomes. Predominantly focused on Independent and Identically Distributed (IID) scenarios, traditional research in MPM often struggles to generalize in Out of Distribution (OOD) scenarios, which are vital for accurate mineral exploration. Addressing these challenges, we introduce an innovative automated conformalized causal learning system for MPM. This system integrates a comprehensive data preprocessing pipeline that includes interpolation, feature filtering, data augmentation, and splitting, effectively managing diverse and imbalanced geological datasets. A central component of the system is Bayesian Optimization, autonomously selecting optimal machine learning models and hyperparameters to significantly enhance performance over non-automated methods. The system's most significant innovation is the incorporation of conformalized causal learning, exceptionally effective in handling OOD data scenarios. This methodology introduces an 'uncertainty region' in predictive models through conformal prediction, substantially reducing misclassification risks, while causal learning elucidates complex cause-and-effect relationships among geological features, essential for precise mineral deposit predictions. We evaluated the performance of our approach on six datasets, where the Area Under the Receiver Operating Characteristic (AUC ROC) of our automated optimized system surpassed the baseline method by an overall 17.84%, and the false positive rate (FPR) was reduced by an overall 84.31%. This development marks a significant advancement in MPM, enhancing accuracy and efficiency in mineral resource exploration and setting a new benchmark in the field. Released as an open-source platform, it offers the geological community a highly efficient, adaptable, and user-friendly tool, poised to revolutionize mineral prospectivity mapping in varied real-world scenarios.

Mineral systems are composed of many interacting components that lead to complex, singular and rare properties of geo-data. In mineral prospectivity mapping (MPM), supervised machine learning algorithms, which have advantages in dealing with complex geo-data, usually involve uncertainty resulting from the discretization of continuous evidential maps into arbitrary classes as well as the large data imbalance caused by the rarity of deposit locations. Consequently, the predicted results may be biased. In this paper, a random forest (RF) algorithm based on the bagging technique is used to map the prospectivity of tungsten polymetallic deposits in the Nanling metallogenic belt. Data-driven logistic transformation is employed to obtain continuous evidential maps. Both discretized and continuous evidential maps are used to generate prospectivity models for comparison. To reduce the data imbalance, the under-sampling method and the synthetic minority over-sampling technique (SMOTE) are implemented to generate balanced datasets. The receiver operating characteristic (ROC) curve and improved prediction-area (P-A) plot are applied to evaluate the prospectivity models. The predictive results show that when using the RF algorithm in MPM, the application of continuous evidential maps can improve the performance of prospectivity models and reduce the uncertainty resulting from the discretization of evidential maps. The prospectivity model trained with a balanced SMOTE-generated dataset shows the best overall performance for improving the percentage of deposit locations that are correctly predicted and decreasing the percentage of non-deposit locations that are inaccurately identified as deposit locations to some extent. In addition, the improved P-A plot is superior to the ROC curve because the latter neglects the occupied area, which is critical for mineral exploration and may provide an overly optimistic performance with imbalanced data. However, further testing of the evaluation criteria and the SMOTE approach to reduce data imbalance is warranted to determine fully the universality in MPM from the perspective of data imbalance. Based on prospectivity models, four high-potential areas and five moderate-potential areas are delineated, which indicates good future prospecting for tungsten polymetallic deposits in this region.

Summary The prime focus of mineral prospectivity mapping (MPM) is the identification, delineation and level of potential for an area to host feasible mineral resources. In general, GIS-based MPM can be either data driven or knowledge-driven. Methods of data-driven MPM, which involve quantitative analysis of spatial relationships among anomalies (i.e., indicators of mineralization) and existing occurrences of mineral deposits of the type sought, is suitable for "brownfields" or well-explored regions, wherein the objective is to define additional targets for exploration. One of the main objectives of WP3 in the FRAME project (www.frame.lneg.pt) is to produce a map of Strategic and Critical Raw Materials (SCRM) for Europe. Another objective is to produce predictive targeting based on GIS exploration tools at continental scale. in this presentation we will review the state of the art for the MPM and show example of the favourability maps using CBA method and hybride fuzzy weight of evidence produced in the FRAME project. This methodology represents different development stages, scales and progress of economic geology surveys which could be a tool to improve effectiveness and efficiency of future investments in exploration. Improved understanding and support for a balance management of competitive land-usage interests is an additional benefit.

Porphyry-type mineral prospectivity mapping with imbalanced data via prior geological transfer learning

Sensitivity analysis of prospectivity modeling to evidence maps: Enhancing success of targeting for epithermal gold, Takab district, NW Iran

Mineral prospectivity mapping (MPM) is a powerful approach for identifying mineralization zones with high potential for economically viable mineral deposits. This study proposes a hybrid framework combining a Wasserstein Generative Adversarial Network with Gradient Penalty (WGAN-GP), a Convolutional Neural Network (CNN) and a Fuzzy-Kernel Extreme Learning Machine (FKELM) to address the challenges of imbalanced and uncertain datasets in mineral exploration. The approach was applied to the G.R. Halli gold prospect, in the Chitradurga Schist Belt, Western Dharwar Craton, India, using nine geochemical pathfinder elements. WGAN-GP generated high-quality negative samples, balancing the dataset and reducing overfitting. Compared with Support Vector Machines, Gradient Boosting, and a baseline CNN, FKELM (AUC = 0.976, accuracy = 92%) and WGAN-GP + CNN (AUC = 0.973, accuracy = 91%) showed superior performance and produced geologically coherent prospectivity maps. Promising gold targets were delineated, closely aligned with known mineralized zones and geochemical anomalies. This hybrid framework provides a robust, cost-effective, and scalable MPM solution for structurally controlled geological tracts, insufficient data terrains, and integration with additional geoscience datasets for other complex mineral systems.

Data-driven logistic-based weighting of geochemical and geological evidence layers in mineral prospectivity mapping

Application of fuzzy AHP method to IOCG prospectivity mapping: A case study in Taherabad prospecting area, eastern Iran

This study applies machine learning (ML) techniques to classify fertile [for porphyry Cu and (or) Au systems] and barren adakites using geochemical data from New Brunswick, Canada. It emphasizes that not all intrusive units, including adakites, are inherently fertile and should not be directly used as the heat source evidence layer in mineral prospectivity mapping without prior analysis. Adakites play a crucial role in mineral exploration by helping distinguish between fertile and barren intrusive units, which significantly influence ore-forming processes. A dataset of 99 fertile and 66 barren adakites was analyzed using seven ML models: support vector machine (SVM), neural network, random forest (RF), decision tree, AdaBoost, gradient boosting, and logistic regression. These models were applied to classify 829 adakite samples from around the world into fertile and barren categories, with performance evaluated using area under the curve (AUC), classification accuracy, F1 score, precision, recall, and Matthews correlation coefficient (MCC). SVM achieved the highest performance (AUC = 0.91), followed by gradient boosting (0.90) and RF (0.89). For model validation, 160 globally recognized fertile adakites were selected from the dataset based on well-documented fertility characteristics. Among the tested models, SVM demonstrated the highest classification accuracy (93.75%), underscoring its effectiveness in distinguishing fertile from barren adakites for mineral prospectivity mapping. Statistical analysis and feature selection identified middle rare earth elements (REEs), including Gd and Dy, with Hf, as key indicators of fertility. A comprehensive analysis of 1596 scatter plots, generated from 57 geochemical variables, was conducted using linear discriminant analysis (LDA) to determine the most effective variable pairs for distinguishing fertile and barren adakites. The most informative scatter plots featured element vs. element combinations (e.g., Ga vs. Dy, Ga vs. Gd, and Pr vs. Gd), followed by element vs. major oxide (e.g., Fe2O3T vs. Gd and Al2O3 vs. Hf) and ratio vs. element (e.g., La/Sm vs. Gd, Rb/Sr vs. Hf) plots, whereas major oxide vs. major oxide, ratio vs. ratio, and major oxide vs. ratio plots had limited discriminatory power.

In mineral exploration, proposing new prospects is a main issue and geospatial information system (GIS) offers different techniques for this procedure. Mineral prospectivity maps (MPMs) can reduce costs and errors for predicting mineral potentials. In this research, data layers have been analyzed and integrated for porphyry copper mineral prospectivity mapping in Chahargonbad sheet of Kerman Province, Central Iran. First of all, geological, geochemical, geophysical, and satellite imagery have been processed separately and assigned according to their favorability in porphyry copper mineralization using index overlay and fuzzy logic methods which was lead to introduce two priorities for further exploration. Eight hundred forty-six geochemical stream sediment samples were analyzed using multifractal concentration-number model to generate geochemical anomaly maps. Afterward, airborne magnetic data was used to extract magnetic edges which can represent structures (faults and fractures) and lithological contacts. Airborne radiometric data is also used to identify felsic units based on K, Th, and U (RGB) ternary image. Thereafter, satellite imagery (ASTER, ETM+ images) was processed based on relative absorption band depth (RBD), principle component analysis (PCA) and Crosta methods in order to separate alteration zones in the study area. Totally, 80 km2 has been detected as the first priority that is located in the northwest (Kuh Panj), center (south of Hamid Abad and east of Sorkhgino), southeast (Chenar Kaf), and southwest (north of Alimoradi) of this sheet. Also, the second priority (97 km2) is located in the center (southeast of Syloieh), southeast (Chenar Kaf), east (Cheheltan), and northwest (south of Kuh Panj).

Predictive modeling of mineral prospectivity by deep self-organizing map: Implications for copper exploration targeting in saveh district, central Iran

An improved data-driven fuzzy mineral prospectivity mapping procedure; cosine amplitude-based similarity approach to delineate exploration targets