Three Essays on Application of Semiparametric Regression: Partially Linear Mixed Effects Model and Index Model

Three Essays on Application of Semiparametric Regression: Partially Linear Mixed Effects Model and Index Model

Drei Aufsätze über Anwendung der Semiparametrischen Regression: Teilweise Lineares Gemischtes Modell und Index Modell

Ren Ohinata

Doctoral thesis

Date of Examination: 2012-05-03

Date of issue: 2012-12-07

Advisor: Sperlich, Stefan Prof. Dr.

Referee: Sperlich, Stefan Prof. Dr.

Referee: Klasen, Stephan Prof. Dr.

Referee: Kneib, Thomas Prof. Dr.

Persistent Address: http://hdl.handle.net/11858/00-1735-0000-000D-F0A2-0

Files in this item

Name: ohinata.pdf

Size: 2,0 MB

Format: PDF

View/Open

The following license files are associated with this item:

Creative Commons

Abstract

English

The dissertation consists of three essays. The main focus is set on semiparametric regression modeling, which embodies the strength of parametric and nonparametric regression models in terms of flexibility, dimensionality and interpretability. The first essay is “Some Recent Advances in Modeling with Mixed Effects for Small Areas, Multi-level and Panel Models

Keywords: Semiparametric model; Kernel regression; Index model; Mixed model; Panel data; Cross validation; Wild bootstrap; Bandwidth selection; Dimension reduction; Principal components; Welfare indicator

Schlagwörter: Semiparametrisches Modell; Kernregression; Index Modell; Gemischtes Modell; Paneldaten; Kreuzvalidierung; Wild Bootstrap; Bandweitenwahl; Dimensionsreduktion; Hauptkomponenten; Wohlfahrtsindikator