Dynamic cluster analysis of educated and smart society development in European Union countries

Abstract

PURPOSE: The aim of the paper is to analyze the educational and smart society development process in the European Union countries, in 2003-2017.

DESIGN/METHODOLOGY/APPROACH: 17 variables have been used for the multivariate analysis of the problem. Multiple regression was the main method for missing data imputation. The number of clusters was identified in dendrogram of Ward’s agglomerative clustering method, and final partition obtained by k-means method. Composite Index of Educational and Smart Development has been proposed to measure the general level of each cluster.

FINDINGS: Five cluster have been identified and characterized. Their dynamic geographical composition changed over time with a tendency for many countries to move toward higher level clusters.

PRACTICAL IMPLICATIONS: It seems that educational part of Europe 2020 Strategy works rather well for most of the EU countries.

ORIGINALITY/VALUE: The choice of variables is always somehow subjective. Dynamic cluster analysis seems to be promising approach in identifying changes in both level and structure. The new measure for cluster stability has been proposed in the paper.