Jul. 7 2009 2:00PM

Jul. 7 2009 3:30PM

The talk is divided in two parts. The first part has an introductory character in order to describe the prediction theory under the general linear model in finite populations, the basic linear mixed models in small area estimation and the poverty indicators of interest. The second part is devoted to describe the area-level time models developed in the European project SAMPLE (http://www.sample-project.eu/). In addition to theoretical considerations, the results of two simulations experiments are presented. The experiments are designed to investigate when it is worthwhile to introduce auto-correlation parameters. An application to Spanish Living Condition Survey data is also presented. The scope of the application is to estimate poverty indicators by sex in Spanish provinces (small areas). Finally, some conclusions are given.

Pravidelný seminář Rozhodování a řízení za neurčitosti. (Contact person: Miroslav Kárný)