3.4 Hierarchical cluster analysis
HCA is an unsupervised classification method, which finishes classification by grouping samples of similar properties. In the present study, two variables were involved into the HCA analysis. One group included major fatty acids with more than 0.5% level together with all minor components; the other one was consist of the structure of phenolic compounds. The dendrograms constructed by standardized data are presented in Fig.4.
As can be seen in Fig.4 (a), Coratina and Koroneiki oils were completely divided into two categories. In addition, the oils from 2-year-old and 11-year-old trees were separated from the oils from 7-year-old trees for both Coratina and Koroneiki. The result demonstrated that the chemical compositions of olive oil were highly cultivar-dependent while tree age had a certain influence on the characteristics of the same varieties. As shown in Fig.4 (b), the oils from 2-year-old Coratina and Koroneiki trees were classified as a group. The structure of phenolic compounds in oil from 7-year-old Koroneiki trees was similar with that from 11-year-old Koroneiki trees. When the 5 distance threshold was chosen, 7-year-old Coratina oil was grouped in the same cluster with 7-year-old and 11-year-old Koroneiki oils. 11-year-old Coratina oil was divided into this cluster at 7 distance threshold. The results implied that both cultivar and tree age have apparent effects the structure of phenolic compounds.