3.5. Principal component analysis
Principal component analysis allowed us to visualize and sum- marize all the differences highlighted by the ANOVA results. The eigenvalues of the correlation matrix showed that the first three Principal Components (PCs) explained 83% of the total variance. The first two PCs explained 41.09% and 29.67%, respectively. The PCA biplot (Fig. 5) showed that total phenols and antioxidant capacity were highly correlated reciprocally and with the same positive di- rection as PC1, being located to the right in the plot. They were responsible for most of the variance captured by PC1. Sodium content and, especially, h were located to the left in the plot, inversely correlated with PC1 and with total phenols and antioxi- dant capacity. On the same side, but with less influence on the PC1
spanning variation, were L* and b* (Fig. 5). Examining the product
distribution (Fig. 5), we find raw purple carrot and purple mild jam
located on the right, corresponding to total phenols and antioxidant capacity, while raw yellow carrot and yellow common jam were located on the left. These results, confirming the ANOVA results, highlighted that purple carrots are rich in antioxidants and phenols and that these traits are not greatly modified by the mild jam processing method.