A data screening approach to confirming a target mineral is chlorite using <scp>EPMA</scp> and <scp>LA‐ICPMS</scp> data

Applying machine learning techniques to large datasets of in situ analyses has been proven be a powerful tool Earth Sciences. However, problems may arise when dealing with minerals such as chlorite, that exist solid solution rather than single, stoichiometric ideal. It can difficult determine whether the variations major element concentrations are due compositional difference mineral interest or sampling surrounding phases addition during analyses. If latter, interpretations results would complicated, misled even spurious. Here we present method identify chlorite based on and minor content, from both LA-ICPMS EPMA data. Further dataset 3,317 have shown 7.4% these include significant quantities non-chlorite material.