Regresyjne badania efektywności produkcji w warunkach niejednorodności zbioru obiektów

Abstract

The article attempts at discriminating a set of objects in the regressive studyin situation of their heterogenity. The first stage consists of a division of the set into typological groups and next, of regressive modelling in the received sub-sets. The following phase, which is in the focus of this study, involves proceeding with group regressive examinations. Suggested approaches to the regressive modelling, distinct to the classical one, is compromising between a general regression related to the whole set and a group regression which is designed in a traditional way i.e. a two phase regression. It proves correct to estimate a general regression for the whole set of objects and for groups and on these grounds — to create a two stage model which can explain changes of an explained variable in the respective groups which are resulting from the properties of the whole set of objects as well as of specific groups. The study is illustrated with the construction of one variable models on the example of simulational data for 30 plants isolated into three typological groups.