Zmienne syntetyczne w modelowaniu ekonometrycznym

Abstract

The present article is an attempt to construct econometric models with synthetic variables, created on the grounds of multidimensional comparative analysis. The author discusses a problem of creating these variables having time-sectional data provided, and next presents a proposition of defining an optimal sub-set of synthetic describing variables utilizing the so-called two- stage method of capacity of information. Information facilitating construction of new weighted synthetic variable is also provided by this method, weights being measures of stock of information of fractional diagnostic variables. The article is concluded with a review of various types of model propositions which can be applied in comparative analysis and in forecasting of socioeconomic development of all analysed economic objects or their groups.