Multiplicative Scatter Correction (Geladi et al., 1985), as its name indicates, is typically used to compensate for light scattering effects and changes in path length. Indeed, light scattering from solid samples and emulsions can result in multiplicative deviations that are dependent on the wavelength (Heise & Winzen, 1985). MSC minimizes these deviations by fitting a linear model between a reference spectrum and other spectra of the dataset using the linear least squares method. The reference spectrum is often chosen as the average of all spectra in the dataset. The model coefficients are used to compute the ideal (i.e. msc corrected) spectra. Following application of MSC, all spectra appear to have the same absorbance level. MSC and SNV usually lead to similar results. However, as mentioned, MSC requires the use of a reference spectrum while SNV does not.
MSC can be performed using the msc function in the pls package:
library(pls)
#Perform MSC correction:
newspectra<-msc(as.matrix(mydata$NIR))
mydataMSC<-data.frame(mydata[,1:6], NIR = I(newspectra))
#Plot new spectra:
myBasicPlot(mydataMSC, wavelengths, xlim, ylim = c(0,0.8)) 