“…NAP/CLS is related to other multivariate regression methods such as partial least-squares (PLS) and principal component regression (PCR), which have been previously applied to food analysis with bore-magnet NMR instruments (Gerbanowski, Rutledge, Feinberg, & Ducauze, 1997;Pedersen, Munck, & Engelsen, 2000;Prestes et al, 2007;Rudi, Guthausen, Burk, Reh, & Isengard, 2008). Although it is PLS that dominates the applications of multivariate regression of NMR data (Micklander et al, 2003), we have chosen NAP/CLS as it outperforms PLS when the number of calibration samples is reduced (Goicoechea & Olivieri, 2001b). It has the advantage of an intuitive basis that enables more feasible estimation of figures of merit (Lorber, 1986;Lorber, Faber, & Kowalski, 1997).…”