摘要:In this study, the first and last names of people (FN and LN), enterprises (EN) (with plants’ species roots in their names) and phytotoponyms (PT) in five Croatian regions are analyzed, in their relationships. The goals of the study were: to determine the correlations between FN, LN, EN and PT; to determine the latent structure of these variables; to forecast number of PT (criterion) on the base of predictors (FN, LN, EN); to determine grouping of the places (within certain regions) as cases by two plants’ categorizations; to determine grouping of the plants as cases by regions. We have analyzed 15 places, grouped in five regions, with 39 different plant species. The results revealed that the only principal component highly positively correlated with the variables last name and office name, while the projections for the variables first name (moderate high) and phytotoponyms (low size) were negative. Prediction of the criteria phytotoponyms is satisfactorily good, using three predictors: last name, first name and the office name. First cluster analysis revealed that phytotoponyms are mostly related with trees and deciduous plants, while names are related with trees, deciduous and herbaceous plants. Second cluster analysis obtained clear distinction between regions in dominant PTs, based on certain plants’ names. The results indicate clear association between phytotoponyms and names of people.
关键词:correlation; nonlinear methods; plant species; regions