文章基本信息

标题：Variablenselektion bei gebundener Hochrechnung
本地全文：下载
作者：Melanie KNOBELSPIES ; Ralf MÜNNICH
期刊名称：Austrian Journal of Statistics
出版年度：2008
卷号：37
期号：3&4
页码：335-335
出版社：Austrian Statistical Society
摘要：In survey sampling, the adequate use of auxiliary variables may considerably improve the efficiency of survey estimates gained from generalized regression techniques. In some surveys like the German income and expenditure survey or the German survey of income and living conditions (D-SILC), a huge amount of potential candidates of auxiliary information is available. Due to methodological and numerical limitations, efficient variable selection need to be applied for gaining efficient estimates. Within this paper, classical statistical variable selection procedures are studied in order to elaborate their efficiency for survey estimation problems. Special emphasis is put on optimizing the model for regression estimation techniques. Additionally, the influence of stratification and allocation on the results will be considered.
关键词：Verallgemeinerter Regressionssch¡§atzer, Hilfsinformationen, Hauptkomponentenregression, Schichtung.