Multiplicative signal correction algorithm

Spectral filters The mathematical treatments of NIR spectra prior to regression used to remove variance that only adds to the error term in the calibration model, (i.e., variance not containing systematic information on the sample). Common spectral filters are multiplicative signal correction (MSC), derivatives and smoothing algorithms. 

Multiplicative signal correction, also known as multiplicative scatter correction (Martens and Naes), was developed to correct for light scattering in reflectance measurements (or measurements containing a strong multiplicative component). Offset and scaling are performed for this algorithm (Fig. 5). 

This resonant Mie scattering effect is the cause of the phenomenon previously referred to as the dispersion artefact , which was an inaccurate description because the effect was not solely due to dispersion and it is not an artefact because it can be explained fully. With an understanding of the physics behind resonant Mie scattering, preliminary algorithms based on the extended multiplicative signal correction (EMSC) have been developed. The EMSC... 

Quite simply, the FCPC algorithm could not correctly manage a scenario where there were multiple spikes in either AOAl or AOA2 that were exactly 1.2 seconds apart (i.e., exactly the length of time that the FCPC would memorize old data in the event of a signal spike). The Australian Transport Safety Board report notes that it... 

---