The discrimination of tectonic setting Using trace elements in zircons： A machine learning approach

Abstract

Abstract Zircon is the most important accessory mineral in geological research, and they record information on isotopes and trace elements which is of great significance in earth science research. Trace elements in Zircons can be used for analyzing the genesis of zircons, calculating the magma temperature and oxygen fugacity, and tracing the magma source. Due to the limitation of visual dimensions, the information on the zircons is mainly shown with the method of low dimensional diagrams in the present studies, so the high dimensional relationships during trace elements of the zircons are difficult to be discovered. However, with the development of machine learning, mining the high dimensional relationships during the trace elements of the zircons becomes possible. In this paper, four supervised learning algorithms including Random Forest, Support Vector Machine, Decision Tree, and eXtreme Gradient Boosting have been implemented to analyze trace elements of 3907 magmatic zircons from the GEOROC database, and a precise 13-dimensional data classifier model has been established in order to distinguish the tectonic settings of the rift, ocean island, and convergent margin. Based on the results of accuracy, precision, recall, and F1-score, the machine learning approach of eXtreme Gradient Boosting is best in the paper and the results of Accuracy, Precision, Recall, and F1-score are 0.948, 0.941, 0.922, 0.930, respectively. In summary, eXtreme Gradient Boosting in the paper could provide a high-dimensional discriminative approach to distinguish the tectonic settings.