Chinautla-Sacojito data

Table U. Principal Component Coefficients for the Chinautla-Sacojito Data Set ...

Figure 6. Principal components plot (components Nos. 1 and 2) based on Chinautla—Sacojito data only.

Table III. Principal Component Coefficients Chinautla- Sacojito Data with Outliners Removed ...

After removal of size from the data set (outliers removed previously), hierarchical and k-means cluster analyses fail to distinguish the two texture groups in the Chinautla-Sacojito data. Discriminant analysis with the size-corrected data yields nonsignificant test statistics for the effect of group (tempered vs. nontempered) and misclassifies 22 of 76 observations. 

Analysis. Two separate patterns should be discernible in the artificial data set sherds, clays, and hypothetical mixtures from Chinautla-Sacojito should be separable from raw materials and products from Durazno and fine-and medium-paste subgroups should be recognizable within the major groups. The first pattern involves shape variation, that is, differing mean vectors and differing variance-covariance structures that arise through nat-... 

---