Data-Driven Predictive Modelling of Mineral Prospectivity Using Machine Learning and Deep Learning Methods: A Case Study from Southern Jiangxi Province, China

Predictive modelling of mineral prospectivity, a critical, but challenging procedure for delineation of undiscovered prospective targets in mineral exploration, has been spurred by recent advancements of spatial modelling techniques and machine learning algorithms. In this study, a set of machine learning methods, including random forest (RF), support vector machine (SVM), artificial neural network (ANN), and a deep learning convolutional neural network (CNN), were employed to conduct a data-driven W prospectivity modelling of the southern Jiangxi Province, China. A total of 118 known W occurrences derived from long-term exploration of this brownfield area and eight evidential layers of multi-source geoscience information related to W mineralization constituted the input datasets. This provided a data-rich foundation for training machine learning models. The optimal configuration of model parameters was trained by a grid search procedure and validated by 10-fold cross-validation. The resulting predictive models were comprehensively assessed by a confusion matrix, receiver operating characteristic curve, and success-rate curve. The modelling results indicate that the CNN model achieves the best classification performance with an accuracy of 92.38%, followed by the RF model (87.62%). In contrast, the RF model outperforms the rest of ML models in overall predictive performance and predictive efficiency. This is characterized by the highest value of area under the curve and the steepest slope of success-rate curve. The RF model was chosen as the optimal model for mineral prospectivity in this region as it is the best predictor. The prospective zones delineated by the prospectivity map occupy 9% of the study area and capture 66.95% of the known mineral occurrences. The geological interpretation of the model reveals that previously neglected Mn anomalies are significant indicators. This implies that enrichment of ore-forming material in the host rocks may play an important role in the formation process of wolframite and can represent an innovative exploration criterion for further exploration in this area.

[1]  Feng Yuan,et al.  3D computational simulation-based mineral prospectivity modeling for exploration for concealed Fe–Cu skarn-type mineralization within the Yueshan orefield, Anqing district, Anhui Province, China , 2019, Ore Geology Reviews.

[2]  V. Rodriguez-Galiano,et al.  Machine learning predictive models for mineral prospectivity: an evaluation of neural networks, random forest, regression trees and support vector machines , 2015 .

[3]  Gongzhuang Peng,et al.  Landslide Susceptibility Assessment Using Integrated Deep Learning Algorithm along the China-Nepal Highway , 2018, Sensors.

[4]  J. Harris,et al.  Comparison of the Data-Driven Random Forests Model and a Knowledge-Driven Method for Mineral Prospectivity Mapping: A Case Study for Gold Deposits Around the Huritz Group and Nueltin Suite, Nunavut, Canada , 2016, Natural Resources Research.

[5]  B. Pradhan,et al.  GIS-based modeling of rainfall-induced landslides using data mining-based functional trees classifier with AdaBoost, Bagging, and MultiBoost ensemble frameworks , 2016, Environmental Earth Sciences.

[6]  Jian‐tang Peng,et al.  Metal source and wolframite precipitation process at the Xihuashan tungsten deposit, South China: Insights from mineralogy, fluid inclusion and stable isotope , 2019, Ore Geology Reviews.

[7]  E. Carranza,et al.  Introduction to the Special Issue: GIS-based mineral potential modelling and geological data analyses for mineral exploration , 2015 .

[8]  Ribana Roscher,et al.  Foreword to the Special Issue on Machine Learning for Geospatial Data Analysis , 2018, ISPRS Int. J. Geo Inf..

[9]  M. Yousefi,et al.  Introduction to the special issue: GIS-based mineral potential targeting , 2017 .

[10]  A. Du,et al.  SHRIMP zircon U–Pb and molybdenite Re–Os isotopic dating of the tungsten deposits in the Tianmenshan–Hongtaoling W–Sn orefield, southern Jiangxi Province, China, and geological implications , 2011 .

[11]  Tongfei Li,et al.  Prospectivity Mapping for Tungsten Polymetallic Mineral Resources, Nanling Metallogenic Belt, South China: Use of Random Forest Algorithm from a Perspective of Data Imbalance , 2019, Natural Resources Research.

[12]  Emmanuel John M. Carranza,et al.  Geometric average of spatial evidence data layers: A GIS-based multi-criteria decision-making approach to mineral prospectivity mapping , 2015, Comput. Geosci..

[13]  L. S. Davis,et al.  An assessment of support vector machines for land cover classi(cid:142) cation , 2002 .

[14]  Thomas Blaschke,et al.  Evaluation of Different Machine Learning Methods and Deep-Learning Convolutional Neural Networks for Landslide Detection , 2019, Remote. Sens..

[15]  Emmanuel John M. Carranza,et al.  Objective selection of suitable unit cell size in data-driven modeling of mineral prospectivity , 2009, Comput. Geosci..

[16]  Christopher J. C. Burges,et al.  A Tutorial on Support Vector Machines for Pattern Recognition , 1998, Data Mining and Knowledge Discovery.

[17]  Qiuming Cheng,et al.  Application of a hybrid method combining multilevel fuzzy comprehensive evaluation with asymmetric fuzzy relation analysis to mapping prospectivity , 2009 .

[18]  Jian Wang,et al.  Deep learning and its application in geochemical mapping , 2019, Earth-Science Reviews.

[19]  Victor F. Rodriguez-Galiano,et al.  Predictive modelling of gold potential with the integration of multisource information based on random forest: a case study on the Rodalquilar area, Southern Spain , 2014, Int. J. Geogr. Inf. Sci..

[20]  Jian‐tang Peng,et al.  Tracing the origin of ore-forming fluids in the Piaotang tungsten deposit, South China: Constraints from in-situ analyses of wolframite and individual fluid inclusion , 2019, Ore Geology Reviews.

[21]  Tomislav Hengl,et al.  Finding the right pixel size , 2006, Comput. Geosci..

[22]  Mao Jingwen,et al.  Major types and time–space distribution of Mesozoic ore deposits in South China and their geodynamic settings , 2013, Mineralium Deposita.

[23]  Nannan Zhang,et al.  Back-propagation neural network and support vector machines for gold mineral prospectivity mapping in the Hatu region, Xinjiang, China , 2018, Earth Science Informatics.

[24]  A. Hezarkhani,et al.  Application of support vector machine for the separation of mineralised zones in the Takht-e-Gonbad porphyry deposit, SE Iran , 2018, Journal of African Earth Sciences.

[25]  Renguang Zuo,et al.  Support vector machine: A tool for mapping mineral prospectivity , 2011, Comput. Geosci..

[26]  Bashir Shokouh Saljoughi,et al.  A comparative analysis of artificial neural network (ANN), wavelet neural network (WNN), and support vector machine (SVM) data-driven models to mineral potential mapping for copper mineralizations in the Shahr-e-Babak region, Kerman, Iran , 2018, Applied Geomatics.

[27]  Yongzhong Zhang,et al.  Geophysical investigations of the geology and structure of the Pangushan-Tieshanlong tungsten ore field, South Jiangxi, China—Evidence for site-selection of the 2000-m Nanling Scientific Drilling Project (SP-NLSD-2) , 2015 .

[28]  Q. Cheng,et al.  Fuzzy Weights of Evidence Method and Its Application in Mineral Potential Mapping , 1999 .

[29]  Vesa Nykänen,et al.  Optimizing a Knowledge-driven Prospectivity Model for Gold Deposits Within Peräpohja Belt, Northern Finland , 2017, Natural Resources Research.

[30]  Frank van Ruitenbeek,et al.  Knowledge-guided data-driven evidential belief modeling of mineral prospectivity in Cabo de Gata, SE Spain , 2008, Int. J. Appl. Earth Obs. Geoinformation.

[31]  M. Yousefi,et al.  Data-driven logistic-based weighting of geochemical and geological evidence layers in mineral prospectivity mapping , 2016 .

[32]  R. Zuo,et al.  Effects of misclassification costs on mapping mineral prospectivity , 2017 .

[33]  Jian‐tang Peng,et al.  In-situ elemental and isotopic compositions of apatite and zircon from the Shuikoushan and Xihuashan granitic plutons: Implication for Jurassic granitoid-related Cu-Pb-Zn and W mineralization in the Nanling Range, South China , 2018 .

[34]  Emmanuel John M. Carranza,et al.  Random forest predictive modeling of mineral prospectivity with small number of prospects and data with missing values in Abra (Philippines) , 2015, Comput. Geosci..

[35]  Yaozu Qin,et al.  Quantitative 3D Association of Geological Factors and Geophysical Fields with Mineralization and Its Significance for Ore Prediction: An Example from Anqing Orefield, China , 2018, Minerals.

[36]  T. Sun,et al.  Joint Application of Fractal Analysis and Weights-of-Evidence Method for Revealing the Geological Controls on Regional-Scale Tungsten Mineralization in Southern Jiangxi Province, China , 2017 .

[37]  Binbin Li,et al.  Application of a Maximum Entropy Model for Mineral Prospectivity Maps , 2019, Minerals.

[38]  Kiyun Yu,et al.  Machine Learning Classification of Buildings for Map Generalization , 2017, ISPRS Int. J. Geo Inf..

[39]  Xie Xuejing,et al.  Geochemical mapping in China , 1997 .

[40]  T. Dawson,et al.  Selecting thresholds of occurrence in the prediction of species distributions , 2005 .

[41]  Amir Hossein Alavi,et al.  Machine learning in geosciences and remote sensing , 2016 .

[42]  Dimitris Kiritsis,et al.  Deep learning for big data applications in CAD and PLM - Research review, opportunities and case study , 2018, Comput. Ind..

[43]  E. Carranza,et al.  Logistic regression for geologically constrained mapping of gold potential, Baguio district, Philippines , 2001 .

[44]  Ufuk Çelik,et al.  The Prediction of Precious Metal Prices via Artificial Neural Network by Using RapidMiner , 2017 .

[45]  Rahul Ramachandran,et al.  Detection of transverse cirrus bands in satellite imagery using deep learning , 2018, Comput. Geosci..

[46]  Sunil Kumar Tripathy,et al.  Performance prediction of gravity concentrator by using artificial neural network-a case study , 2014 .

[47]  A. Hezarkhani,et al.  The use of decision tree induction and artificial neural networks for recognizing the geochemical distribution patterns of LREE in the Choghart deposit, Central Iran , 2017 .

[48]  Yi Wang,et al.  Comparison of convolutional neural networks for landslide susceptibility mapping in Yanshan County, China. , 2019, The Science of the total environment.

[49]  Mei-Fu Zhou,et al.  Genetic types, mineralization styles, and geodynamic settings of Mesozoic tungsten deposits in South China , 2017 .

[50]  E. Carranza,et al.  Data-driven predictive mapping of gold prospectivity, Baguio district, Philippines: Application of Random Forests algorithm , 2015 .

[51]  E. Carranza,et al.  Mapping mineral prospectivity through big data analytics and a deep learning algorithm , 2018, Ore Geology Reviews.

[52]  D. Singer,et al.  Classification of mineral deposits into types using mineralogy with a probabilistic neural network , 1997 .

[53]  V. Nykänen,et al.  Receiver operating characteristics (ROC) as validation tool for prospectivity models — A magmatic Ni–Cu case study from the Central Lapland Greenstone Belt, Northern Finland , 2015 .

[54]  Martin Hale,et al.  A predictive GIS model for mapping potential gold and base metal mineralization in Takab area, Iran , 2001 .

[55]  Masoud Shariat Panahi,et al.  The application of median indicator kriging and neural network in modeling mixed population in an iron ore deposit , 2011, Comput. Geosci..

[56]  W. Liu,et al.  GIS-based mineral prospectivity mapping using machine learning methods: A case study from Tongling ore district, eastern China , 2019, Ore Geology Reviews.

[57]  Biswajeet Pradhan,et al.  Spatial prediction models for shallow landslide hazards: a comparative assessment of the efficacy of support vector machines, artificial neural networks, kernel logistic regression, and logistic model tree , 2016, Landslides.

[58]  Feng Chengyou,et al.  Chronology of the Tungsten Deposits in Southern Jiangxi Province, and Episodes and Zonation of the Regional W‐Sn Mineralization – Evidence from High‐precision Zircon U‐Pb, Molybdenite Re‐Os and Muscovite Ar‐Ar Ages , 2012 .

[59]  Lyudmila Sukhostat,et al.  Lithological facies classification using deep convolutional neural network , 2019, Journal of Petroleum Science and Engineering.

[60]  Emmanuel John M. Carranza,et al.  Artificial Neural Networks for Mineral-Potential Mapping: A Case Study from Aravalli Province, Western India , 2003 .

[61]  Geoffrey E. Hinton,et al.  Deep Learning , 2015, Nature.

[62]  R. Zuo,et al.  A comparative study of fuzzy weights of evidence and random forests for mapping mineral prospectivity for skarn-type Fe deposits in the southwestern Fujian metallogenic belt, China , 2016, Science China Earth Sciences.

[63]  R. Zuo.  Machine Learning of Mineralization-Related Geochemical Anomalies: A Review of Potential Methods , 2017, Natural Resources Research.

[64]  Tom Gedeon,et al.  Artificial neural networks: A new method for mineral prospectivity mapping , 2000 .

[65]  R. Vieira,et al.  Fluid-rock interaction is decisive for the formation of tungsten deposits , 2017 .

[66]  Carlos Roberto de Souza Filho,et al.  Probabilistic neural networks applied to mineral potential mapping for platinum group elements in the Serra Leste region, Carajás Mineral Province, Brazil , 2009, Comput. Geosci..

[67]  Yongliang Chen,et al.  A Bat-Optimized One-Class Support Vector Machine for Mineral Prospectivity Mapping , 2019, Minerals.

[68]  Feng Yuan,et al.  Three-dimensional mineral prospectivity modeling for targeting of concealed mineralization within the Zhonggu iron orefield, Ningwu Basin, China , 2015 .

[69]  M. Hale,et al.  Bayesian network classifiers for mineral potential mapping , 2006, Comput. Geosci..

[70]  Erik Cambria,et al.  Recent Trends in Deep Learning Based Natural Language Processing , 2017, IEEE Comput. Intell. Mag..

[71]  E. Carranza,et al.  Selection of coherent deposit-type locations and their application in data-driven mineral prospectivity mapping , 2008 .

[72]  Leo Breiman,et al.  Random Forests , 2001, Machine Learning.