Dear all, I noticed an paper using "machine learning" method to calcualte ferric iron in biotite. This "big data" is composed of several hundreds of biotite analyses by EMPA, with the Fe3+/ΣFe ratios determined by Mössbauer spectroscopy or micro-XANES, and the H2O contents were quantified by FTIR,
SIMS or other independent methods. Principal components regression is one of simplest data analysis methods. Do you think the result is sound for biotite analysis? How much impact would this method of exploring EMPA data will bring to EMPA-related research?
We are a community of analysts, that cares about EPMA