Four essays on statistical problems of hedonic methods : an application to single-family houses
X + 100 p.
Thèse de doctorat: Université de Fribourg, 2013
English
Over the last ten years, the hedonic approach has been acknowledged as the most appropriate method for addressing the valuation of goods that have a non-constant quality. This thesis is structured in four independent papers that investigate statistical problems related to this approach. The aim is to improve actual knowledge in the hedonic field through an empirical or theoretical approach, and provide results that are useful for researchers and practitioners. The price index problem in the hedonic context is discussed in the first two papers. The first paper introduces a theoretical framework for estimating hedonic price indices and their confidence intervals. The second paper analyzes the asymptotic properties of the most common hedonic price indices. The third paper focuses on variable selection in hedonic models where multicollinearity is present. In particular, it proposes a new variable selection algorithm that outperforms ordinary automated selection techniques. The fourth paper implements a methodology for comparing the prediction accuracy of two hedonic models.
-
Faculty
- Faculté des sciences économiques et sociales et du management
-
Language
-
-
Classification
-
Economics
-
Notes
-
- Ressource en ligne consultée le 25.02.2014
-
License
-
License undefined
-
Identifiers
-
-
Persistent URL
-
https://folia.unifr.ch/unifr/documents/303506