Zastosowanie analizy skupień w estymacji regresyjnej dla małych obszarów

Abstract

Information about the whole population or its part are used in the regression estimation of small area parameters. In the paper the possibilities of application of cluster analysis methods are considered in case of determining the group of similar small areas. The studies of a similarity of subpopulations are conducted on the basis of studies of similarity of regression function and similarity of ranks for small areas. The results of simulation analysis of precision of regression estimators are presented in case of using two auxiliary variables.