These scripts use mid-infrared spectroscopy data to predict soil properties from a library of wet chemistry results.
The scripts run a PLS and RF algorithm on spectra, the code is augmented with Monte Carlo error estimation and a generalisable code base (to allow more models to be tested)