Dataset and R-codes for Publication: "Best performances of visible-near infrared models in soils with little carbonate - a field study in Switzerland" (accepted version) (original) (raw)

Published February 22, 2024 | Version 1.0

Dataset Open

Description

In this upload you can find the R-codes and dataset for the Publication:

"Best performances of visible-near infrared models in soils with little carbonate - a field study in Switzerland" by Simon Oberholzer Laura Summerauer, Markus Steffens and Chinwe Ifejika Speranza accepted in SOIL (https://doi.org/10.5194/egusphere-2023-1087)

To reproduce the results of the manuscript, start with the R-file “ResampleandPreprocess.R” to prepare spectral data and then continue with the R-file “PLSRmodelling_GroupedCV.R” for the modelling.

The R-files “control_train.R” and “rep_grouped_kfold_CV.R” are helper-functions for the grouped cross-validation. The file metadata.csv explains the column names in the spectral data (spcdata.RDS).

Files

Zenodo_upload.zip

Files (21.1 MB)