Mapping mineral prospectivity through big data analytics and a deep learning algorithm

• Published in: Ore geology reviews Vol. 102; no. C; pp. 811 - 817
• Main Authors: Xiong, Yihui, Zuo, Renguang, Carranza, Emmanuel John M.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 01.11.2018; Elsevier
• Subjects: Big data; Deep learning; GIS; Mapping mineral prospectivity; Deep learning; GIS; Big data; Mapping mineral prospectivity
• ISSN: 0169-1368; 1872-7360
• DOI: 10.1016/j.oregeorev.2018.10.006

[Display omitted] •Big data analytics and a deep learning algorithm were used to map mineral prospectivity.•42 geological, geochemical and geophysical variables were involved.•A case from southwestern Fujian metalorganic zone of China was presented. Identification of anomalies related to mineralization and integration of multi-source geoscience data are essential for mapping mineral prospectivity. In this study, we applied big data analytics and a deep learning algorithm to process geoscience data to identify and integrate anomalies related to skarn-type Iron mineralization in the southwestern Fujian metallogenic zone of China. Based on the geological setting and environment for the formation of skarn-type Iron mineralization, 42 relevant variables, including two geological, one geophysical, and 39 geochemical variables, were analyzed and integrated for detecting anomalies related to mineralization using a deep autoencoder network. The results indicate that the mapped prospectivity areas have a strong spatial relationship with the locations of known mineralization and demonstrate that big data analytics supported by deep learning methods is a potential technique to be considered for use in mineral prospectivity mapping.