Model-Based Pre-Processing in Vibrational Spectroscopy
Achim Kohler, Johanne Heitmann Solheim, Valeria Tafintseva, Boris Zimmermann, and Volha Shapaval
Pre-processing of vibrational spectroscopic data is generally facilitating subsequent data modeling and spectral interpretation. While different approaches are in use for pre-processing spectral data, infrared and Raman spectroscopies have made extensive use of model-based pre-processing by extended multiplicative signal correction. In this chapter, we introduce the concept of model-based pre-processing and discuss different situations where extended multiplicative signal correction (EMSC) has been used to remove instrumental effects, scatter variations and other interferent effects from absorbance spectra. We discuss EMSC and other approaches such as taking derivatives, Standard Normal Variate (SNV), Vector Normalization and Peak normalization for pre-processing.