repositoryrepositoryClassic scatter-corrected PLS — Multiplicative Scatter Correction with a 10-component PLS regressor.
unvalidatedThe classic near-infrared regression pipeline.
spectrum to remove scatter from particle size and path-length differences.
The default first choice for powdered or granular samples where scatter dominates. MSC and SNV are close cousins — try both; MSC is preferred when a stable reference spectrum is meaningful for your sample set.
{
"name": "MSC + PLS",
"pipeline": [
{"class": "nirs4all.operators.transforms.MultiplicativeScatterCorrection"},
{"class": "sklearn.preprocessing.StandardScaler"},
{"y_processing": {"class": "sklearn.preprocessing.StandardScaler"}},
{"class": "sklearn.model_selection.KFold", "params": {"n_splits": 5, "shuffle": true, "random_state": 7}},
{"model": {"class": "sklearn.cross_decomposition.PLSRegression", "params": {"n_components": 10}}, "name": "PLS-10"}
]
}
# Python
import nirs4all_repository as n4r
pipe = n4r.get("msc_pls")
config = pipe.to_nirs4all() # ready for nirs4all.run() / predict()# any language: read the index, fetch + verify
curl https://repository.nirs4all.org/data/index.json
curl https://repository.nirs4all.org/data/pipelines/msc_pls/pipeline.json