“…Spectral preprocessing techniques that have been applied include (but not limited to) the following: (1) averaging replicate spectra, (2) maximum normalisation, (3) fitting a weighted (inverse measurement variance) smoothing spline to each spectra with direct extraction of smoothed reflectance, (4) use of first and second derivatives of the original spectra, (5) individually combining the resulting average reflectance with the first-and second-derivatives, (6) multiplicative and extended multiplicative scatter correction, (7) Savitzky-Golay, and (8) standard normal variate Dematte et al, 2004;Viscarra-Rossel et al, 2006b;Westad et al, 2008;Yu et al, 2008;Sundaram et al, 2010;Chen et al, 2010;Kuang and Mouazen, 2011). Alternatively, variableselection techniques, which minimize the number of simultaneous variables to be used in modelling, have also been applied in this regard.…”