A data grouping was observed from the scatter plot of the PC
scores (Figure 4a), but the data for samples adulterated at low
concentration appear to overlap more. This confirmed our
assumption that the different spectral attributes among the
samples were associated with the concentration of the starch in
the onion powder. Furthermore, we assessed the potential of
PCA to distinguish among the low starch concentration
samples (1−10%), but the data were overly scattered and
overlapping (data not shown). The first three PC loadings
account for nearly 99% of the total variance in the samples. PC
loadings are correlation coefficients between the PC scores and
the original variables, which measure the importance of each
variable in accounting for the variability in the PC. It is possible
to interpret the first few PCs in terms of overall effect of a
contrast between groups of variables based on the structures of
PC loadings. The PC loading plot (Figure 4b) shows the
highest loading for PC1 and PC3 at ∼1960 nm, related to the
starch. PC1 and PC2 show other important bands between
1435 and 1450 nm, corresponding to the first O−H stretching
overtone, also assigned to the starch.