Spectral Preprocessing Methods Reference

Spectral Preprocessing Methods

A reference guide to common preprocessing techniques used in chemometric analysis of spectral data.

Standard Normal Variate (SNV)

Centers and scales each spectrum individually. Useful for correcting scatter effects.

Savitzky-Golay Derivatives

Smooths data and computes derivatives. First derivative removes baseline offset; second derivative removes linear baseline drift.

Parameters: window size, polynomial order, derivative order

Multiplicative Scatter Correction (MSC)

Corrects for path length differences and scatter effects by regressing each spectrum against a reference spectrum.

Mean Centering

Subtracts the mean spectrum. Standard first step in most chemometric analyses.

Autoscaling

Mean centers and divides by standard deviation. Gives equal weight to all variables.