A machine learning-based approach to exploration targeting of porphyry Cu-Au deposits in the Dehsalm district, eastern Iran

The Dehsalm district in the central part of the Lut block, is an important polymetallic belt of eastern Iran. This study identifies the prospective areas for Cu-Au porphyry deposits within the Dehsalm district by integrating the outputs of the Restricted Boltzmann Machine (RBM) and Random Forest (RF) models. Based on geological, metallogenic and statistical analyses, seven predictor maps, namely proximity to faults and intrusive rocks, to aeromagnetics anomalies, to hydrothermal alterations (argillic and Fe-oxide) and two multivariate geochemical signatures of the stream sediments, were generated to train the models. To identify anomalies indicative of Cu-Au porphyry mineralization from the RBM results, the RF model was applied as a supervised machine-learning algorithm. The target areas delineated in the combined model occupy 4.2% of the study region and contain 82% of the known Cu-Au occurrences. The most prospective areas identified in the combined model extend along with structures that are spatially associated with known Cu-Au occurrences, and provide clear targets for future exploration.