Primena multivarijacione analize u formiranju kompozitnog indeksa razvijenosti okruga u Srbiji

Abstract

One of the most important issues of economic policy of all countries is procuring a uniform economic development on entire territory. The regional unbalance of the degree of development among some territories in the Republic of Serbia constitutes one of the most complex development problems. Accentuated population decrease in some areas on one side, and large population and economy concentration in just a few cities on the other, produce negative effects on economic, social, spatial and ecological domain of the development of every country. One of the segments of this study will focus on the choice of indicators used for determining the degree of regional development and for comprehending the factors that determine a regional development. For the purpose of analyzing the differences in development of territorial units in the Republic of Serbia, in this particular case – the districts, a methodology for quantification of regional differences, from the aspect of a number of relevant criteria, was developed by compacting individual indicators into a composite index of development. The application of multivariate data analysis is of great importance in establishing the composite index, for the reason of getting an insight of correlations among variables (chosen indicators). It could include correlation analysis, regression analysis, factor analysis and others. In this dissertation, the emphasis will be placed on the factor analysis. The purpose of applying this method of multivariate analysis is that of determining whether different information, contained in the chosen district development indicators, is sufficiently balanced within the composite index. When aggregating indicators into a composite index, some weights (indicator weights) obtained based on factor analysis results will be added. The goal is that the weight of indicator weights corresponds to the relative importance of each variable (indicator) within the composite index. In this particular case, indicator weights would be the adequate factorial loading values obtained from factor analysis. After calculating the composite index for each district and its posterior analysis, the final result of this doctoral thesis would be the ranking of districts in the Republic of Serbia according to the values of the obtained composite index of development.