NIR spectral data were collected using a Luminar 5030 Miniature “Hand held” AOTFNIR (Acousto-Optic Tunable Filter-Near Infrared) Analyser (Brimrose Corporation of America, Baltimore, Maryland) in reflectance mode. A spectral range of 1100–2300 nm with 601 points (2 nm step) was used to obtain the spectra at room temperature. Each sample was scanned at four different points and the average spectrum was used for subsequent analysis. Each spectrum was an average of 50 scans. Data were then transformed to absorbance for further analysis. Analyses First, an analysis of the two varieties was performed in order to identify undesirable scatter effects. With this purpose in mind, absorbance data of both measurements (peeled and unpeeled) of each potato variety were examined to identify imperfections and to compare the degree of scatter between varieties. Then, the mean absorbance spectrum of both peeled and unpeeled tubers of each variety was obtained to identify the main differences in absorption at different wavelengths in the NIR spectrum. Finally, and with the aim of evaluating whether the differences graphically observed corresponded to the formation of statistically different groups, a classification was performed using support vector machines (SVM). Prior to SVM classification, samples were pre-treated using SNV to reduce the physical variability between the samples as a consequence of the scat