Mineral Systems Analysis and Artificial Neural Network Modeling of Chromite Prospectivity in the Western Limb of the Bushveld Complex,South Africa

Radial basis function link neural network (RBFLN) and fuzzy-weights of evidence (fuzzy-WofE) methods were used to assess regional-scale prospectivity for chromite deposits in the Western Limb and the Nietverdiend layered mafic intrusion of the Bushveld Complex in South Africa. Five predictor maps derived from geological, geochemical and geophysical data were processed in a GIS environment and used as spatial proxy for critical processes that were most probably responsible for the formation of the chromite deposits in the study area. The RBFLN was trained using input feature vectors that correspond to known deposits, prospects and non-deposits. The training was initiated by varying the number of radial basis functions (RBFs) and iterations. The results of training the RBFLN provided optimum number of RBFs and iterations that were used for classification of the input feature vectors. The results show that the network classified 73% of the validation deposits into highly prospective areas for chromite deposit, covering 6.5% of the study area. The RBFLN entirely classified all the non-deposit validation points into low prospectivity areas, occupying 86.6% of the study area. In general, the efficiency of the RBFLN in classifying the validation deposits and non-deposits indicates the degree of spatial relationship between the input feature vectors and the training points, which represent chrome mines and prospects. The RBFLN and fuzzy-WofE analyses used in this study are important in guiding identification of regional-scale prospect areas where further chromite exploration can be carried out.     

Knowledge-Driven and Data-Driven Fuzzy Models for Predictive Mineral Potential Mapping

In this paper, we describe new fuzzy models for predictive mineral potential mapping: (1) a knowledge-driven fuzzy model that uses a logistic membership function for deriving fuzzy membership values of input evidential maps and (2) a data-driven model, which uses a piecewise linear function based on quantified spatial associations between a set of evidential evidence features and a set of known mineral deposits for deriving fuzzy membership values of input evidential maps. We also describe a graphical defuzzification procedure for the interpretation of output fuzzy favorability maps. The models are demonstrated for mapping base metal deposit potential in an area in the south-central part of the Aravalli metallogenic province in the state of Rajasthan, western India. The data-driven and knowledge-driven models described in this paper predict potentially mineralized zones, which occupy less than 10% of the study area and contain at least 83% of the model and validation base metal deposits. A cross-validation of the favorability map derived from using one of the models with the favorability map derived from using the other model indicates a remarkable similarity in their results. Both models therefore are useful for predicting favorable zones to guide further exploration work.     

Mineral Favorability Mapping: A Comparison of Artificial Neural Networks,Logistic Regression,and Discriminant Analysis

A Probabilistic Neural Network (PNN) was trained to classify mineralized and nonmineralized cells using eight geological, geochemical, and geophysical variables. When applied to a second (validation) set of well-explored cells that had been excluded from the training set, the trained PNN generalized well, giving true positive percentages of 86.7 and 93.8 for the mineralized and nonmineralized cells, respectively. All artifical neural networks and statistical models were analyzed and compared by the percentages of mineralized cells and barren cells that would be retained and rejected correctly respectively, for specified cutoff probabilities for mineralization. For example, a cutoff probability for mineralization of 0.5 applied to the PNN probabilities would have retained correctly 87.66% of the mineralized cells and correctly rejected 93.25% of the barren cells of the validation set. Nonparametric discriminant analysis, based upon the Epanechnikov Kernel performed better than logistic regression or parametric discriminant analysis. Moreover, it generalized well to the validation set of well-explored cells, particularly to those cells that were mineralized. However, PNN performed better overall than nonparametric discriminant analysis in that it achieved higher percentages of correct retention and correct rejection of mineralized and barren cells, respectively. Although the generalized regression neural network (GRNN) is not designed for a binary—presence or absence of mineralization— dependent variable, it also performed well in mapping favorability by an index valued on the interval [0, 1]. However, PNN outperformed GRNN in correctly retaining mineralized cells and rejecting barren cells of the validation set.     

Radial Basis Functional Link Nets Used as a Prospectivity Mapping Tool for Orogenic Gold Deposits Within the Central Lapland Greenstone Belt, Northern Fennoscandian Shield

Among the more popular spatial modeling techniques, artificial neural networks (ANN) are tools that can deal with non-linear relationships, can classify unknown data into categories by using known examples for training, and can deal with uncertainty; characteristics that provide new possibilities for data exploration. Radial basis functional link nets (RBFLN), a form of ANN, are applied to generate a series of prospectivity maps for orogenic gold deposits within the Paleoproterozoic Central Lapland Greenstone Belt, Northern Fennoscandian Shield, Finland, which is considered highly prospective yet clearly under explored. The supervised RBFLN performs better than previously applied statistical weights-of-evidence or conceptual fuzzy logic methods, and equal to logistic regression method, when applied to the same geophysical and geochemical data layers that are proxies for conceptual geological controls. By weighting the training feature vectors in terms of the size of the gold deposits, the classification of the neural network results provides an improved prediction of the distribution of the more important deposits/occurrences. Thus, ANN, more specifically RBFLN, potentially provide a better tool to other methodologies in the development of prospectivity maps for mineral deposits, hence aiding conceptual exploration.     

Application of Artificial Neural Network for Gold–Silver Deposits Potential Mapping: A Case Study of Korea

The aim of this study is to analyze hydrothermal gold–silver mineral deposits potential in the Taebaeksan mineralized district, Korea, using an artificial neural network (ANN) and a geographic information system (GIS) environment. A spatial database considering 46 Au and Ag deposits, geophysical, geological, and geochemical data was constructed for the study area using the GIS. The geospatial factors were used with the ANN to analyze mineral potential. The Au and Ag mineral deposits were randomly divided into a training set (70%) to analyze mineral potential using ANN and a test set (30%) to validate predicted potential map. Four different training datasets determined from likelihood ratio and weight of evidence models were applied to analyze and validate the effect of training. Then, the mineral potential index (MPI) was calculated using the trained back-propagation weights, and mineral potential maps (MPMs) were constructed from GIS data for the four training cases. The MPMs were then validated by comparison with the test mineral occurrences. The validation results gave respective accuracies of 73.06, 73.52, 70.11, and 73.10% for the training cases. The comparison results of some training cases showed less sensitive to training data from likelihood ratio than weight of evidence. Overall, the training cases selected from 10% area with low and high index value of MPM_L and MPM_W gave higher accuracy (73.52 and 73.10%) for MPMs than those (73.06 and 70.11%, respectively) from known deposits and 10% area with low index value of MPI_L and MPI_W.     

Application of Radial Basis Functional Link Networks to Exploration for Proterozoic Mineral Deposits in Central Iran

The metallogeny of Central Iran is characterized mainly by the presence of several iron, apatite, and uranium deposits of Proterozoic age. Radial Basis Function Link Networks (RBFLN) were used as a data-driven method for GIS-based predictive mapping of Proterozoic mineralization in this area. To generate the input data for RBFLN, the evidential maps comprising stratigraphic, structural, geophysical, and geochemical data were used. Fifty-eight deposits and 58 ‘nondeposits’ were used to train the network. The operations for the application of neural networks employed in this study involve both multiclass and binary representation of evidential maps. Running RBFLN on different input data showed that an increase in the number of evidential maps and classes leads to a larger classification sum of squared error (SSE). As a whole, an increase in the number of iterations resulted in the improvement of training SSE. The results of applying RBFLN showed that a successful classification depends on the existence of spatially well distributed deposits and nondeposits throughout the study area. An erratum to this article can be found at     

A MaxEnt Model for Mineral Prospectivity Mapping

Mineral prospectivity mapping is an important preliminary step for mineral resource exploration. It has been widely applied to distinguish areas of high potential to host mineral deposits and to minimize the financial risks associated with decision making in mineral industry. In the present study, a maximum entropy (MaxEnt) model was applied to investigate its potential for mineral prospectivity analysis. A case study from the Nanling tungsten polymetallic metallogenic belt, South China, was used to evaluate its performance. In order to deal with model over-fitting, varying levels of β _j -regularization were set to determine suitable β value based on response curves and receiver operating characteristic (ROC) curves, as well as via visual inspections of prospectivity maps. The area under the ROC curve (AUC = 0.863) suggests good performance of the MaxEnt model under the condition of balancing model complexity and generality. The relative importance of ore-controlling factors and their relationships with known deposits were examined by jackknife analysis and response curves. Prediction–area (P–A) curves were used to determine threshold values for demarcating high probability of tungsten polymetallic deposit occurrence within small exploration area. The final predictive map showed that high favorability zones occupy 14.5% of the study area and contain 85.5% of the known tungsten polymetallic deposits. Our study suggests that the MaxEnt model can be efficiently used to integrate multisource geo-spatial information for mineral prospectivity analysis.     

Evaluation of Weights of Evidence to Predict Epithermal-Gold Deposits in the Great Basin of the Western United States

The weights-of-evidence method provides a simple approach to the integration of diverse geologic information. The application addressed is to construct a model that predicts the locations of epithermal-gold mineral deposits in the Great Basin of the western United States. Weights of evidence is a data-driven method requiring known deposits and occurrences that are used as training sites in the evaluated area. Four hundred and fifteen known hot spring gold–silver, Comstock vein, hot spring mercury, epithermal manganese, and volcanogenic uranium deposits and occurrences in Nevada were used to define an area of 327.4 km² as training sites to develop the model. The model consists of nine weighted-map patterns that are combined to produce a favorability map predicting the distribution of epithermal-gold deposits. Using a measure of the association of training sites with predictor features (or patterns), the patterns can be ranked from best to worst predictors. Based on proximity analysis, the strongest predictor is the area within 8 km of volcanic rocks younger than 43 Ma. Being close to volcanic rocks is not highly weighted, but being far from volcanic rocks causes a strong negative weight. These weights suggest that proximity to volcanic rocks define where deposits do not occur. The second best pattern is the area within 1 km of hydrothermally altered areas. The next best pattern is the area within 1 km of known placer-gold sites. The proximity analysis for gold placers weights this pattern as useful when close to known placer sites, but unimportant where placers do not exist. The remaining patterns are significantly weaker predictors. In order of decreasing correlation, they are: proximity to volcanic vents, proximity to east-west to northwest faults, elevated airborne radiometric uranium, proximity to northwest to west and north-northwest linear features, elevated aeromagnetics, and anomalous geochemistry. This ordering of the patterns is a function of the quality, applicability, and use of the data. The nine-pattern favorability map can be evaluated by comparison with the USGS National Assessment for hot spring gold–silver deposits. The Spearman's ranked correlation coefficient between the favorability and the National Assessment permissive tracts is 0.5. Tabulations of the areas of agreement and disagreement between the two maps show 74% agreement for the Great Basin. The posterior probabilities for 51 significant deposits in the Great Basin, both used and not used in the model, show the following: 26 classified as favorable; 25 classified as permissive; and 1, Florida Canyon, classified as nonpermissive.The Florida Canyon deposit has a low favorability because there are no volcanic rocks near the deposit on the Nevada geologic map used. The largest areas of disagreement are caused by the USGS National Assessment team concluding that volcanic rocks older than 27 Ma in Nevada are not permissive, which was not assumed in this model. The weights-of-evidence model is evaluated as reasonable and delineates permissive areas for epithermal deposits comparable to expert's delineation. The weights-of-evidence model has the additional characteristics that it is well defined, reproducible, objective, and provides a quantitative measure of confidence.     

Application of Data-Driven Evidential Belief Functions to Prospectivity Mapping for Aquamarine-Bearing Pegmatites,Lundazi District,Zambia

A case application of data-driven estimation of evidential belief functions (EBFs) is demonstrated to prospectivity mapping in Lundazi district (eastern Zambia). Spatial data used to represent recognition criteria of prospectivity for aquamarine-bearing pegmatites include mapped granites, mapped faults/fractures, mapped shear zones, and radioelement concentration ratios derived from gridded airborne radiometric data. Data-driven estimates EBFs take into account not only (a) spatial association between an evidential map layer and target deposits but also (b) spatial relationships between classes of evidences in an evidential map layer. Data-driven estimates of EBFs can indicate which spatial data provide positive or negative evidence of prospectivity. Data-driven estimates of EBFs of only spatial data providing positive evidence of prospectivity were integrated according to Dempster’s rule of combination. Map of integrated degrees of belief was used to delineate zones of relative degress of prospectivity for aquamarine-bearing pegmatites. The predictive map has at least 85% prediction rate and at least 79% success rate of delineating training and validation deposits, respectively. The results illustrate usefulness of data-driven estimation of EBFs in GIS-based predictive mapping of mineral prospectivity. The results also show usefulness of EBFs in managing uncertainties associated with evidential maps.     

Application of the Geo-Anomaly Unit Concept in Quantitative Delineation and Assessment of Gold Ore Targets in Western Shandong Uplift Terrain,Eastern China

A number of large and giant ore deposits have been discovered within the relatively small areas of lithospheric structure anomalies, including various boundary zones of tectonic plates. The regions have become the well-known intercontinental ore-forming belts, such as the circum-Pacific gold–copper, copper–molybdenum, and tungsten–tin metallogenic belts. These belts are typical geological anomalous areas. An investigation into the hydrothermal ore deposits in different regions in the former Soviet Union illustrated that the geologic structures of ore fields of almost all major commercial deposits have distinct features compared with the neighboring areas. These areas with distinct features are defined as geo-anomalies. A geo-anomaly refers to such a geologic body or a combination of bodies that their composition, texture–structure, and genesis are significantly different from those of their surroundings. A geo-anomaly unit (GU) is an area containing distinct features that can be delineated with integrated ore-forming information using computer techniques on the basis of the geo-anomaly concept. Herein, the GU concept is illustrated by a case study of delineating the gold ore targets in the western Shandong uplift terrain, eastern China. It includes: (1) analyses of gold ore-forming factors; (2) compilation of normalized regional geochemical map and extraction of geochemical anomalies; (3) compilation of gravitational and aeromagnetic tectonic skeleton map and extraction of gravitational and aeromagnetic anomalies; (4) extraction of circular and linear anomalies from remote-sensing Landsat TM images; (5) establishment of a geo-anomaly conceptual model associated with known gold mineralization; (6) establishment of gold ore-forming favorability by computing techniques; and (7) delineation and assessment of ore-forming units. The units with high favorability are suggested as ore targets.     

A Hybrid Fuzzy Weights-of-Evidence Model for Mineral Potential Mapping

This paper describes a hybrid fuzzy weights-of-evidence (WofE) model for mineral potential mapping that generates fuzzy predictor patterns based on (a) knowledge-based fuzzy membership values and (b) data-based conditional probabilities. The fuzzy membership values are calculated using a knowledge-driven logistic membership function, which provides a framework for treating systemic uncertainty and also facilitates the use of multiclass predictor maps in the modeling procedure. The fuzzy predictor patterns are combined using Bayes’ rule in a log-linear form (under an assumption of conditional independence) to update the prior probability of target deposit-type occurrence in every unique combination of predictor patterns. The hybrid fuzzy WofE model is applied to a regional-scale mapping of base-metal deposit potential in the south-central part of the Aravalli metallogenic province (western India). The output map of fuzzy posterior probabilities of base-metal deposit occurrence is classified subsequently to delineate zones with high-favorability, moderate favorability, and low-favorability for occurrence of base-metal deposits. An analysis of the favorability map indicates (a) significant improvement of probability of base-metal deposit occurrence in the high-favorability and moderate-favorability zones and (b) significant deterioration of probability of base-metal deposit occurrence in the low-favorability zones. The results demonstrate usefulness of the hybrid fuzzy WofE model in representation and in integration of evidential features to map relative potential for mineral deposit occurrence.     

Spatial Modeling for Base-Metal Mineral Exploration Through Integration of Geological Data Sets

This study involves the integration of information interpreted from data sets such as LandsatTM, Airborne magnetic, geochemical, geological, and ground-based data of Rajpura—Dariba,Rajasthan, India through GIS with the help of (1) Bayesian statistics based on the weights ofevidence method and (2) a fuzzy logic algorithm to derive spatial models to target potentialbase-metal mineralized areas for future exploration. Of the 24 layers considered, five layers(graphite mica schist (GMS), calc-silicate marble (CALC), NE-SW lineament 0–2000 mcorridor (L4-NESW), Cu 200–250 ppm, and Pb 200–250 ppm) have been identified from theBayesian approach on the basis of contrast. Thus, unique conditions were formed based onthe presence and absence of these five map patterns, which are converted to estimate posteriorprobabilities. The final map, based on the same data used to determine the relationships, showsfour classes of potential zones of sulfide mineralization on the basis of posterior probability.In the fuzzy set approach, membership functions of the layers such as CALC, GMS, NE-SWlineament corridor maps, Pb, and Cu geochemical maps have been integrated to obtain thefinal potential map showing four classes of favorability index.     

Classification of Soil Groups Using Weights-of-Evidence-Method and RBFLN-Neural Nets

Weights-of-Evidence (WofE) and Radial Basis Function Link Net (RBFLN) were applied to soil group mapping in eastern Finland. The data consisted of low altitude airborne geophysical measurements, Landsat 5 TM-satellite image, and digital elevation model (DEM) and slope information derived from it. Probability maps were constructed for each soil group one by one and combined into a prediction map of soil groups using maximum posterior probability (WofE) or pattern membership (RBFLN). Self-Organizing Map (SOM) and Sammon’s Mapping were applied for selecting the data sets for modeling and visualizing the data. The soil types belonging to each soil group used in the Arc-SDM modeling were defined by clusters revealed by the SOM and Sammon’s Mapping algorithms. The soil types with similar characters were collected in the same cluster. Numerical evaluation of the models’ performance was performed using the confusion matrix. The Ratio of Correct Classifications (RCC) for the best WofE model was 0.64 in the training area and 0.61 in the testing area. The RCC for the best RBFLN model was 0.62. Modeling of soil groups using Arc-SDM is time consuming because models need to be constructed for each soil group before combining them into a final prediction map. In this study a simple method was tested for combining the maps. In the future, more attention should be paid to combining the posterior probability models and also to selecting data sets used for modeling.     

Use of Noise to Augment Training Data: A Neural Network Method of Mineral–Potential Mapping in Regions of Limited Known Deposit Examples

One of the main factors that affects the performance of MLP neural networks trained using the backpropagation algorithm in mineral-potential mapping isthe paucity of deposit relative to barren training patterns. To overcome this problem, random noise is added to the original training patterns in order to create additional synthetic deposit training data. Experiments on the effect of the number of deposits available for training in the Kalgoorlie Terrane orogenic gold province show that both the classification performance of a trained network and the quality of the resultant prospectivity map increasesignificantly with increased numbers of deposit patterns. Experiments are conducted to determine the optimum amount of noise using both uniform and normally distributed random noise. Through the addition of noise to the original deposit training data, the number of deposit training patterns is increased from approximately 50 to 1000. The percentage of correct classifications significantly improves for the independent test set as well as for deposit patterns in the test set. For example, using ±40% uniform random noise, the test-set classification performance increases from 67.9% and 68.0% to 72.8% and 77.1% (for test-set overall and test-set deposit patterns, respectively). Indices for the quality of the resultant prospectivity map, (i.e. D/A, D × (D/A), where D is the percentage of deposits and A is the percentage of the total area for the highest prospectivity map-class, and area under an ROC curve) also increase from 8.2, 105, 0.79 to 17.9, 226, 0.87, respectively. Increasing the size of the training-stop data set results in a further increase in classification performance to 73.5%, 77.4%, 14.7, 296, 0.87 for test-set overall and test-set deposit patterns, D/A, D × (D/A), and area under the ROC curve, respectively.     

Developing a geographic expert system for regional mapping of volcanogenic massive sulfide (VMS) deposit potential

A personal computer-based geographic information system (GIS) is used to develop a geographic expert system (GES) for mapping and evaluating volcanogenic massive sulfide (VMS) deposit potential. The GES consists of an inference network to represent expert knowledge, and a GIS to handle the spatial analysis and mapping. Evidence from input maps is propagated through the inference network, combining information by means of fuzzy logic and Bayesian updating to yield new maps showing evaluation of hypotheses. Maps of evidence and hypotheses are defined on a probability scale between 0 and 1. Evaluation of the final hypothesis results in a mineral potential map, and the various intermediate hypotheses can also be shown in map form.The inference net, with associated parameters for weighting evidence, is based on a VMS deposit model for the Chisel Lake deposit, a producing mine in the Early Protoerzoic Snow Lake greenstone belt of northwest Manitoba. The model is applied to a small area mapped at a scale of 1:15,840. The geological map, showing lithological and alteration units, provides the basic input to the model. Spatial proximity to contacts of various kinds are particularly important. Three types of evidence are considered: stratigraphic, heat source, and alteration. The final product is a map showing the relative favorability for VMS deposits. The model is implemented as aFortran program, interfaced with the GIS. The sensitivity of the model to changes in the parameters is evaluated by comparing predicted areas of elevated potential with the spatial distribution of known VMS occurrences.     

Mineral Prospectivity Prediction from High-Dimensional Geoscientific Data Using a Similarity-Based Density Estimation Model

Assuming a study region in which each cell has associated with it an N-dimensional vector of values corresponding to N predictor variables, one means of predicting the potential of some cell to host mineralization is to estimate, on the basis of historical data, a probability density function that describes the distribution of vectors for cells known to contain deposits. This density estimate can then be employed to predict the mineralization likelihood of other cells in the study region. However, owing to the curse of dimensionality, estimating densities in high-dimensional input spaces is exceedingly difficult, and conventional statistical approaches often break down. This article describes an alternative approach to estimating densities. Inspired by recent work in the area of similarity-based learning, in which input takes the form of a matrix of pairwise similarities between training points, we show how the density of a set of mineralized training examples can be estimated from a graphical representation of those examples using the notion of eigenvector graph centrality. We also show how the likelihood for a test example can be estimated from these data without having to construct a new graph. Application of the technique to the prediction of gold deposits based on 16 predictor variables shows that its predictive performance far exceeds that of conventional density estimation methods, and is slightly better than the performance of a discriminative approach based on multilayer perceptron neural networks.     

An Integrated Mapping Approach to Petroleum Exploration on the Pratt Anticline,Pratt County,Kansas

A quantitative map comparison/integration technique to aid in petroleum exploration was applied to an area in south-central Kansas. The visual comparison and integration of maps has become increasingly difficult with the large number and different types of maps necessary to interpret the geology and assess the petroleum potential of an area; therefore, it is desirable to quantify these relationships. The algebraic algorithm used in this application is based on a point-by-point comparison of any number and type of spatial data represented in map form. Ten geological and geophysical maps were compared and integrated, utilizing data from 900 wells located in a nine-township area on the Pratt Anticline in Pratt County, Kansas. Five structure maps, including top of the Lansing Group (Pennsylvanian), Mississippian chert, Mississippian limestone, Viola Limestone (Ordovician), and Arbuckle Group (Cambro-Ordovician), two isopachous maps from top of Mississippian chert to Viola and Lansing to Arbuckle, a Mississippian chert porosity map, Bouguer gravity map, and an aeromagnetic map were processed and interpreted. Before processing, each map was standardized and assigned a relative degree of importance, depending on knowledge of the geology of the area. Once a combination of weights was obtained that most closely resembled the pattern of proved oil fields (target map), a favorability map was constructed based on a coincidence of similarity values and of geological properties of petroleum reservoirs. The resulting favorability maps for the study area indicate location of likely Mississippian chert and lower Paleozoic production.     

One-level prediction—A numerical method for estimating undiscovered metal endowment

One-level prediction has been developed as a numerical method for estimating undiscovered metal endowment within large areas. The method is based on a presumed relationship between a numerical measure of geologic favorability and the spatial distribution of metal endowment. Metal endowment within an unexplored area for which the favorability measure is greater than a favorability threshold level is estimated to be proportional to the area of that unexplored portion. The constant of proportionality is the ratio of the discovered endowment found within a suitably chosen control region, which has been explored, to the area of that explored region. In addition to the estimate of undiscovered endowment, a measure of the error of the estimate is also calculated. One-level prediction has been used to estimate the undiscovered uranium endowment in the San Juan basin, New Mexico, U.S.A. A subroutine to perform the necessary calculations is included.     

Predictive Mapping of Prospectivity in the Gurupi Orogenic Gold Belt,North–Northeast Brazil: An Example of District-Scale Mineral System Approach to Exploration Targeting

The Gurupi Belt hosts a Paleoproterozoic gold province located in north–northeastern Brazil, at the borders of Pará and Maranhão states. It is considered to be an extension of the prolific West African Craton’s Birimian gold province into South America. Additionally, the belt has been the object of recent mineral exploration programs with significant resource discoveries. This study presents the results of predictive mapping using up-to-date mineral system concepts and recently finished regional-scale geological mapping, stream sediment and airborne geophysical surveys conducted by the Geological Survey of Brazil. We relate gold mineralization to an initially enriched crust, metamorphism, deep fluid pathways, structurally controlled damage zones and hydrothermal alteration. Prospective targets were generated using only regional public datasets and knowledge-driven targeting technique. This work did not incorporate any known gold deposits, yet it predicted the largest known deposits and their satellite targets. Besides, high prospective targets mapped almost 40% of known primary gold occurrences within 7% of the project area. This work allowed considerable search area reduction and identification of new target areas, thus collaborating on reducing costs, time and risk of mineral exploration. Results indicate that we achieved an efficient understanding of the geological processes related to the Gurupi Belt mineral system.     

Knowledge Recovery for Continental-Scale Mineral Exploration by Neural Networks

This study is concerned with understanding of the formation of ore deposits (precious and base metals) and contributes to the exploration and discovery of new occurrences using artificial neural networks. From the different digital data sets available in BRGM's GIS Andes (a comprehensive metallogenic continental-scale Geographic Information System) 25 attributes are identified as known factors or potential factors controlling the formation of gold deposits in the Andes Cordillera. Various multilayer perceptrons were applied to discriminate possible ore deposits from barren sites. Subsequently, because artificial neural networks can be used to construct a revised model for knowledge extraction, the optimal brain damage algorithm by LeCun was applied to order the 25 attributes by their relevance to the classification. The approach demonstrates how neural networks can be used efficiently in a practical problem of mineral exploration, where general domain knowledge alone is insufficient to satisfactorily model the potential controls on deposit formation using the available information in continent-scale information systems.