Machine Learning Prediction of Quartz Forming?Environments

Trace elements of quartz document the physical-chemical evolutions growth, which has been a great and most applied tool in study geological settings quartz-forming environments. A classic method is using graphic diagram plots visualizing trace element discriminations trends, examples including Al-Ti (Rusk, 2012, https://doi.org/10.1007/978-3-642-22161-3_14) Ti-Al-Ge (Schrön et al., 1988, https://www.researchgate.net/publication/236149159_Geochemische_Untersuchungen_an_Pegmatitquarzen). However, those diagrams are limited to two dimensions cannot show information higher dimension. In study, we thus used machine learning-based approach evaluate elements, visualized them for first time high-dimensional diagrams. We revisited 1,626 samples from nine environments previous studies, support vector characterize values contained Al, Ti, Li, Ge, Sr. demonstrate that machines can identify crystallization environment with significantly accuracy than traditional plotting methods. Our work massively improve confidence on distinguishing origin different high efficiency. The may also be applicable other minerals, anticipate our research starting point investigating mineral learning techniques. classifier accessed via https://quartz-classifier.herokuapp.com.