Multivariate analysis can though be applied for other purposes
than just clustering items. It is possible to evaluate the material
conservation state using PCA. Indeed, the Raman spectra fitting
parameters of the external parts of the archeological copals were
considered to PCA (Fig. 7), The variables behavior was then
studied, and specific criteria that sort the samples depending on
their surface degradation state could be extracted. On Fig. 7 we can
observe that the samples are well sorted thanks to PC1. Indeed, on
the negative side of PC1 we have all the red copal samples
whereas on the positive side we have the yellow copal samples.
Thus, PC1 contains a sufficient part of the exploited variance to
separate the two groups of copals. Fig. 7 also describes the
projection of the variables obtained after the PCA calculation.
The full width at half maximum of bands 5 and 10 (W5 and W10)
are correlated variables which have an important contribution to
PC1,and which are significant in terms of structural interpretation.
As seen before, band 10 is assigned to the vibration of CH vinyl
groups and decreases with alteration. So if we plot the items on
the base of these two variables, we obtain a chart that sorts the
samples according to an axis, represented by an arrow on Fig. 8.