An introduction to missing data in statistical applications is given in the beginning. The main part of the book deals with selection models for nonignorable missing data. The theory of selection models is described and illustrated by examples. Maximum Likelihood as well as Bayesian estimation approaches are discussed. A selection model with a nonparametric missing model that allows to treat flexible missing patterns is developed. This approach is unique in literature. The proposed model is extended to a model for longitudinal data.
Reihe
Thesis
Sprache
Verlagsort
Frankfurt a.M.
Deutschland
Zielgruppe
Editions-Typ
Illustrationen
Maße
Höhe: 21 cm
Breite: 14.8 cm
Gewicht
ISBN-13
978-3-631-53499-1 (9783631534991)
Schweitzer Klassifikation
The Author: Sandro Scheid was born in Munich 1969. After his studies of Economics at the University of Bremen and the Freie Universität of Berlin, between 1989 and 1993, the author gained a degree in Statistics at the Department of Statistics at the University of Munich in 2001. He worked with a research project of the German Scientific Foundation that deals with discrete structures. The author's topic within this project was missing data. He finished his doctoral thesis in 2004.
Contents: Introduction to missing data problems in general - Selection models for nonignorable missing data - Discussion of Maximum Likelihood and Bayesian estimation routines - Development of a selection model with a nonparametric missing model - Extension of the latter to a model for longitudinal data.
