ABSTRACT
Principal component regression is sometimes described as “performing a least-squares regression of the projections of the data onto the basis vectors of a factor space using Inverse least-squares.” A common form of artifact removal is baseline correction of a spectrum or chromatogram. Common linearizations are the conversion of spectral transmittance into spectral absorbance and the multiplicative scatter correction for diffuse reflectance spectra. Centering, sometimes called mean centering, is simply the subtraction of the mean absorbance at each wavelength from each spectrum. The main reason for centering data is to prevent data points that are farther from the orgin form exerting an undue amount of leverage over the points that are closer to the origin. There are many possible ways to scale or weight our data. Scaling or weighting involves multiplying all of the spectra by a different scaling factor for each wavelength. Indicator functions are based upon an analysis of either the eigenvalues or the errors.
