Fig. (S4) form supplementary information is shown the canonical
variates for iECVA method based on 20 components PLS model
in the inner relation for selected intervals (intervals number 17 and
18). In Fig. (S5) of supplementary information, the corresponding
canonical weight vectors are represented. The canonical variates
which are shown in Fig. (S4) describe the classification power of
ECVA method. The canonical variates in the first direction clearly
show the discrimination of butter and olive samples from other
groups whereas those in the second direction show the discrimination
of corn, canola and olives from butter and Soya. For example,
those having high negative variates in the first direction and moderate
positive variates in the second direction are most probably
butter and those having positive variates in both directions are
most probably sunflower or those having negative variate in both
directions are most probably olive sample. On the other hand those
samples that their respective canonical variates in the first and second directions are small positive and negative can be considered
that there are most probably belong to corn