Distinguishing IOCG and IOA deposits via random forest algorithm based on magnetite composition

Abstract

Iron oxide-copper-gold (IOCG) and iron oxide-apatite (IOA) are two significant mineral deposit types with similar tectonic settings and hydrothermal alteration characteristics. There are huge differences in the geological setting, alteration system, and ore-forming fluid composition among IOCG and IOA deposits, leading to controversial genesis. Distinguishing between these two deposit types is significant to reveal the origin of IOCG and IOA systems. In this study, random forest (RF) was employed to classify IOCG and IOA deposits based on the chemical composition of magnetite measured by the electron probe microanalyzer (EPMA) and laser-ablation inductively coupled plasma mass spectrometry (LA-ICP-MS). The obtained results show that (1) a relatively high overall classification accuracy (0.76 for EPMA data and 0.91 for LA-ICP-MS) was obtained via the RF, indicating that the elemental composition of magnetite can effectively distinguish IOCG and IOA deposits; (2) the performance of the RF model based on LA-ICP-MS data is better than that of EMPA data, indicating that the application of more geochemical variables is helpful in distinguishing IOCG and IOA deposits; and (3) the elements V, Mg, and Mn in EPMA data, and Si, Mg, and V in LA-ICP-MS data are identified as the key elements for distinguishing IOCG and IOA deposits.