Mineral Potential Modeling of Podiform Chromite Deposits in the South Neyshabur Ophiolitic Belt Using Independent Component Analysis

Independent component analysis (ICA) is a relatively new multivariable statistical method originally devised for the blind source separation (BSS) problem, where there is no information on how to mix primary sources (mixed signals) and only the necessary condition is independence of the primary signals. Hence, ICA can be used in mineral potential modeling where several independent mineralization processes result in observed variables such as geophysical and geochemical information, and we do not know how the geophysical and geochemical effects of different mineralization processes are mixed together. In this study, we tried to introduce the ICA method as a knowledge-driven method of mineral potential modeling. To this end, an area of 4800 square kilometers in south of Neyshabur, northeast of Iran, was investigated to map the mineral potential of podiform chromite deposits. In this regard, geochemical stream sediment sampling data, ophiolitic facies map, structural pattern of fractures and serpentinite alteration location in the region were used for this study. Finally, the results of mineral potential modeling by the ICA method were compared with the results of univariate and multivariate geochemical studies and were also validated by using locations of the known mineral prospects in the region and receiver operating characteristic (ROC) method. As a result, the area under the ROC curve was marked by 0.967, indicating the outstanding performance of the ICA modeling.