3.3.1. Heavy metals
The median values of the 13 heavy metals, F− and CN− present
in the waste of each thermal process were used to perform the multivariate
statistical analysis. Three principal components (PC) with
eigenvalues higher than 1 were extracted by applying a Varimax
rotation with Kaiser Normalization in PCA. Fig. 2a represents the
association between the metals whereas the Fig. 2c represents the
association between the 11 processes. PC1, PC2, and PC3 explained
24%, 24%, and 19% (total of 67%) of the variance in Fig. 2a and 42%,
25% and 20% (total 87%) of variance in Fig. 2c. To confirm the associations,
PCA is often used with the cluster analysis (CA) [45,46].
Square Euclidian distances of standardized median values (Z scores)
were used for clustering by applying the Ward’s method. The hierarchical
clustering were performed and presented as a dendrogram
by applying variables such as heavy metals (Fig. 2b) and thermal
processes (Fig. 2d).